WO1994026163A1 - Method and system to enhance medical signals - Google Patents
Method and system to enhance medical signals Download PDFInfo
- Publication number
- WO1994026163A1 WO1994026163A1 PCT/US1994/004846 US9404846W WO9426163A1 WO 1994026163 A1 WO1994026163 A1 WO 1994026163A1 US 9404846 W US9404846 W US 9404846W WO 9426163 A1 WO9426163 A1 WO 9426163A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- matrix
- compressed
- singular
- medical
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S3/00—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
- G01S3/80—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using ultrasonic, sonic or infrasonic waves
- G01S3/802—Systems for determining direction or deviation from predetermined direction
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/88—Sonar systems specially adapted for specific applications
- G01S15/89—Sonar systems specially adapted for specific applications for mapping or imaging
- G01S15/8906—Short-range imaging systems; Acoustic microscope systems using pulse-echo techniques
- G01S15/899—Combination of imaging systems with ancillary equipment
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/04—Display arrangements
- G01S7/06—Cathode-ray tube displays or other two dimensional or three-dimensional displays
- G01S7/20—Stereoscopic displays; Three-dimensional displays; Pseudo-three-dimensional displays
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/52—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00
- G01S7/52017—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00 particularly adapted to short-range imaging
- G01S7/52053—Display arrangements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/52—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00
- G01S7/56—Display arrangements
- G01S7/62—Cathode-ray tube displays
- G01S7/6245—Stereoscopic displays; Three-dimensional displays; Pseudo-three dimensional displays
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2576/00—Medical imaging apparatus involving image processing or analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y10—TECHNICAL SUBJECTS COVERED BY FORMER USPC
- Y10S—TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y10S128/00—Surgery
- Y10S128/901—Suppression of noise in electric signal
Definitions
- the invention relates to the analysis of various types of medical data signals, including: electrocardiography (ECG) data; electroencephalography (EEG) data; conventional radiographic data; fluoroscopy data; angiography data, including cine-loops (e.g.
- echocardiography data including ultrasound data; phonocardiography; conventional co ⁇ puted tomography (CT) and ultrafast computed tomography (UFCT) data, including cine- loops; magnetic resonance imaging (MRI) data, including multi-slice, 3D analysis; MRI spectroscopy (MRIS) data; positron emission tomography (PET) data; single positron emission computer tomography (SPECT) data; and other data signals obtained using medical monitoring apparati which contain information which is representative of a subject's medical condition.
- These data signals may contain both normal medical characteristics and abnormal medical characteristics. Some of the data of particular interest to the diagnostician may be subtle and difficult to discern from other signal components that are received, along with interference and noise. Numerous types of medical signal data are generated in the normal course of medical test procedures, and the present invention allows these data to be analyzed more effectively than is currently possible.
- the medical diagnostician often may have some prior understanding, or models, of the underlying physical mechanisms, prior history data, or other a priori information which may help guide the interpretation of the new medical data being acquired. Therefore, a data processing system which enables the incorporation of such models, histories and a priori information is necessary to enable the diagnostician to make the best medical diagnosis, based on all pertinent information available.
- the present invention is used with data obtained from different medical modalities to reduce noise and unwanted signals and to enhance signals of interest. For example, it is used with ECG data to enhance the heartbeat data, allowing electrical alternans, ventricular late potentials, and other diagnostically important features in the heartbeat signals to be detected.
- the system of the present invention improves the contrast, reduces noise, and enhances edges of conventional radiographic data. It improves wall- motion displays and enhances 3D cine-loop displays of fluoroscope and angiograph data.
- B- and M-scan data may be correlated, or data fused, and cine-loops may be enhanced, compressed, transmitted and displayed.
- the present invention improves the spectral resolution of both conventional computed tomography (CT) and ultrafast computed tomography (UFCT) data, and allows for rapid “deblurring, " back projection and inversion of the data for image reconstruction, as well as cine-loop enhancement. It improves the resolution, enhances contrast, reduces noise, and enhances cine-loop retrieval and display of magnetic resonance imaging (MRI) data.
- MRI magnetic resonance imaging
- PET positron emission tomography
- SPECT single photon emission computer tomography
- noisy data not only include random, noise ⁇ like components, but may also contain non-random, signal ⁇ like components, such as unwanted features in the data and AC interference.
- signal averaging reduces noise, it does not reduce these non-random, signal-like, components. Unfortunately these unwanted components may obscure and interfere with the recognition of weak signals of interest.
- the multi-dimensional medical processing and display system of the present invention is used with an existing medical monitoring and imaging system to reduce noise and unwanted signals and to enhance signals of interest in medical data.
- the medical system provides data, obtained from one or more sensors, to the processing system. As the data is collected, it is arranged into a matrix.
- the dimensions of the matrix vary depending upon the form of the data, from one dimension, up to any number of multiple dimensions.
- the medical data obtained from a particular medical monitoring device such as an ECG monitor, a phonocardiographic monitor, EEG monitor, conventional radiographic equipment, a fluoroscopic system, angiographic system, echocardiographic system, an ultrasound device, as well as CT, UFCT, MRI, MRIS, PET, or SPECT systems, or other medical devices, could inherently be comprised of one, two, three, or higher-dimensional data.
- the data are reformatted by the present invention into a single twodimensional matrix X. When data of three or higher dimensions are obtained, the data are separated into multiple two-dimensional matrices, which are then concatenated along a common dimension, forming one large two-dimensional matrix.
- data that are inherently one-dimensional When data that are inherently one-dimensional is obtained, they are segmented and reformatted into a two-dimensional matrix.
- multiple medical modalities may be correlated, "data fused", or processed simultaneously, including data fusion of M-scan, B-scan and doppler data in ultrasound; data fusion of ECG and B-scan ultrasound; data fusion of ECG and phonocardiograph data; and data fusion of EEG waveforms and SPECT.
- a historic database and other a priori information such as a compressed form of modelled medical data, or otherwise anticipated results, is concatenated with the two-dimensional matrix.
- These data allow interfering signals and noise in the data to be diminished and other signals of interest to be enhanced.
- the data are then analyzed efficiently using singular value decomposition (SVD) .
- the two-dimensional, concatenated matrix X is decomposed to obtain a compressed and enhanced form of the data matrix.
- the data matrix X is decomposed to obtain its corresponding left and right singular vectors , R and singular values D.
- the singular vectors are partitioned into one or more groups, or subspaces, either on the basis of their corresponding singular values, by evaluating their singular vectors, or by some other criteria.
- singular vectors from one or more subspaces are selected for further processing to enhance features of interest and to remove noise and other unwanted signal components.
- the selected singular vectors are transformed into their Hankel matrix form, concatenated together, then decomposed using singular value decomposition.
- the resulting left singular vectors P are transformed into their frequency domain representations by the Fast Fourier Transform (FFT) , or related spectral techniques, for further analysis.
- the corresponding right singular vectors Q represent features within the original signal which have been adaptively filtered, or otherwise modified, to enhance and display signal components of interest and may be analyzed further as well.
- This analysis enables the results to be classified into subspaces with the intent of retaining and further enhancing those singular vectors associated with features of interest and reducing or eliminating those singular vectors which are associated with interference, noise, or other unwanted components .
- the autocorrelation matrix is calculated by reformatting the singular vectors into their Hankel matrix form H, then calculating the cross- product matrix HH fc (or alternatively B ⁇ H) .
- the autocorrelation matrix of each singular vector is computed by first taking the Fast Fourier Transform (FFT) of the singular vectors, then multiplying the results by their complex conjugate, and then transforming the results back the Inverse FFT (IFFT) .
- FFT Fast Fourier Transform
- IFFT Inverse FFT
- the resulting autocorrelation function is then formatted to its Toeplitz matrix form A by placing multiple copies of the function in a matrix, with each copy offset at equal intervals to the right.
- the autocorrelation matrix ⁇ is calculated, it is decomposed into its corresponding singular vectors P, fc and singular values D.
- the autocorrelation matrix is decomposed using eigenanalysis into its corresponding eigenvalues and eigenvectors .
- the resulting singular vectors P (or eigenvectors) are separated into subspaces .
- Certain of the autocorrelation singular vectors P are selected to filter out signal components of interest from singular vectors L, Rt, which represent unwanted components, such as AC interference and noise.
- the singular values D of the autocorrelation matrix are modified and used to adjust the weights of the associated singular vectors P.
- the weighted singular vectors are then combined and the resulting coefficients are used as a Finite Impulse Response (FIR) filter to enhance the original singular vectors and/ or R into enhanced singular values L e , R e .
- FIR Finite Impulse Response
- Enhanced medical data containing features of interest X e is then generated from the enhanced original matrix singular values L e and R e , and modified singular values D e , and the results may be displayed to the diagnostician in various formats.
- e , R e and/or X e may be thresholded to improve subsequent visualization and data analysis.
- the enhanced data X e is then displayed.
- the enhanced data is rotated and displayed from different perspectives.
- the rotation is efficiently achieved by first rotating the singular vectors to the appropriate angle, then expanding only that portion of the full dataset which will be visible. With this rotation of the compressed singular vectors, rather than the full, threedimensional dataset, the amount of required computations are substantially reduced.
- Fig. A shows the enhancer system of the present invention in use with a medical monitoring system
- Fig. 1 is a flow chart for the main algorithm of the preferred embodiment of the present invention.
- Fig. 2 is a flow chart of the Preprocessing Function
- Fig. 3 is the history database for features of interest, features not of interest, and model-based features in the preferred embodiment
- Fig. 4 is a flow chart of the Subspace Selection Function
- Fig. 5a is a flow chart of an alternative embodiment of the Subspace Enhancement Function
- Fig. 5b is a flow chart of the preferred embodiment of the Subspace Enhancement Function
- Fig. 6 is a graphic display of a singular vector
- Fig. 7 is the Hankel matrix form of the singular vector of Fig. 6;
- Fig. 8 shows concatentated Hankel matrices
- Fig. 9a is a singular vector
- Fig. 9b is a graphic display of Ps, the FIR filter coefficients
- Fig. 9c is the Subspace-Enhanced form of the singular vector of Fig. 9a;
- Fig. 10 is a flow chart of the Signal Processing Function
- Fig. 11 is a flow chart of the Subspace Expansion Function.
- Figs. 12a-12d show rotation and display of the three ⁇ dimensional data cube through rotation of singular vectors .
- the enhancer system 2 of the present invention improves the quality of medical data that is collected from a patient 6 using a medical monitoring system 8.
- the enhancer system 2 either may be built into a medical monitor or may be added to an existing medical monitor 8.
- the enhancer 2 reduces or eliminates unwanted data, such as background noise, AC interference, effects of respiration, and unwanted dominant features, and also enhances various features in the medical data, which are displayed to the diagnostician on display 4.
- the present invention may be used with medical data obtained from various types of medical monitoring devices, and is not limited to any particular device.
- the system reduces noise and enhances signals of interest in the particular data being analyzed.
- a medical diagnostician can focus in on features of interest in the ECG data such as electrical alternans, ventricular late potentials, or other important features.
- phonocardiographic monitoring of the heart it may be used to isolate and localize abnormalities in the heart valve function, including stenosis, mitral valve prolapse, etc. It isolates and enhances the waveform structures in the data and improves signal localization in EEG data.
- the system improves contrast resolution, edge enhancement, noise reduction and the display of conventional radiographic data.
- the system improves wall motion displays, 3D cine-loop motion enhancement and other displays of fluoroscopic and angiographic data; it correlates or fuses B- and M-scan echocardiographic data, and enhances the cine-loop, data compression, transmission and display of the data.
- It data fuses ECG and phonocardiographic data.
- the system improves the spectral resolution, image reconstruction, projection and rotation, as well as cine-loop enhancement, of conventional CT and UFCT data. It improves the resolution, enhances the contrast, reduces noise, and enhances the cine- loop retrieval in the display of MRI data.
- the present invention improves the spectral resolution of MRIS data as well as improving the resolution, enhancing the contrast, and aiding in image reconstruction using PET and SPECT data. It enables the simultaneous correlation of EEG waveforms and SPECT data.
- a flow chart for the main algorithm 10 of the present invention is shown.
- medical input data 11 is received from a medical monitoring system, and is premultiplied, weighted and reformatted into a single two-dimensional matrix by Preprocessing Function 12.
- the Subspace Selection Function 13 compresses and refines the medical data in the two-dimensional data matrix.
- the data matrix is compressed by decomposing it into its singular values D and singular vectors L and R.
- the data are refined by modifying or eliminating some of the singular values in D. In this manner, unwanted features, interference, and noise are suppressed and/or eliminated from the data.
- the Subspace Selection Function selects a subset of the compressed and refined data for further processing.
- Subsequent processing is performed on the subset of singular vectors, rather than on the entire set of raw data, resulting in a substantial reduction in the overall processing load of the subsequent stages of processing.
- the selected subset in the form of refined left and right singular vectors L, R fc , and singular values D are passed from the Subspace Selection Function 13 to the Subspace Enhancement Function 14 for further processing.
- the Subspace Enhancement Function 14 converts each selected singular vector L, and/or Rt into an autocorrelation matrix A.
- the autocorrelation matrix A is decomposed into singular vectors P and singular values E. Because A is symmetric, the left and right singular vectors are equivalent.
- the results are partitioned into subspaces, and selected subspaces are used to filter the original singular vectors L and/or Rt to further eliminate undesirable features and further refine features of interest.
- the Signal Processing Function 15 analyzes selected singular vectors e , R fc e to further enhance signal components of interest.
- the left and right singular vectors L e and Rt e which represent the adaptively filtered waveforms, may also be modified to enhance and display signal components of interest for example, through thresholding.
- the singular values D within each of the groups of singular vectors L e , R fc e are also modified to form D e , to enhance certain features of interest in the data and to diminish other, unwanted features within the data.
- the Subspace Expansion Function 16 generates an enhanced data matrix X e from e , D e and R fc e .
- history data 18 is simultaneously passed backwards from the Subspace Selection Function 13, the Subspace Enhancement Function and/or the Signal Processing Function to the Preprocessing Function 12 to assist in data refinement and monitoring features of interest.
- Fig. 2 shows the Preprocessing Function 12 in greater detail.
- the Preprocessing Function 12 premultiplies, weights and reformats the medical input data received from a medical monitoring system and concatenates the reformatted data with history data and model based data.
- the input data is prefiltered at 21 by premultiplying the input data by selected history data Lf n , Rf n representing features not of interest in the data. Prefiltering removes unwanted signal components and thereby emphasizes other signal components of interest.
- the history data is in the form of selected singular vectors which have been passed back from the Subspace Selection Function 13, the Subspace Enhancement Function 14 and the Signal Processing Function 15.
- the history data singular vectors representing data not of interest Lf n , Rf n are used to remove selected features not of interest from the input data, such as artifacts due to respiration. For example, new medical data received at time 1 (Xt:l) is premultiplied by the transpose of L e fn» then expanded by premultiplying by e f n and subtracting the result from t . i. yielding Xr
- Xr t:1 Xr t:1 (I - R efn Rt efn ) . In this manner, features not of interest are suppressed or removed from the input data in an initial pass by the Preprocessing Function.
- the Preprocessing Function weights the input data at 22 by predefined weights 25 to accentuate or suppress raw data at certain locations in the raw data matrix.
- Each data value has a weight value for each dimension. For example, in three-dimensional data such as ultrasound, each data value has three weight coefficients. The first coefficient ⁇ scales the data in the x dimension, the second Yj scales the data in the y dimension, and coefficient z j ⁇ scales it in the z-dimension. Similarly, one dimensional data has one vector of associated coefficients, two dimensional data has two vectors of associated coefficients, and higher dimensional data has the appropriate number of additional vectors of coefficients required to scale the data along the other dimensions.
- weight coefficients can either be preset, set by an automatic function, or interactively altered by the diagnostician.
- a diagnostician examining ultrasound data may be particularly interested in the data that corresponds to an organ or structure located a certain distance into the body and not interested in data that corresponds to distances beyond the structure.
- the diagnostician would indicate that the weights corresponding to the distance of particular interest should be increased, and all weights corresponding to distances on either side of the structure should be decreased. The system would then set the distance weights accordingly.
- the input data is reformatted from its original format into a two-dimensional matrix (shown at 23) .
- the medical data such as ECG, phonocardiograms, EEG, conventional radiography, fluoroscopy, angiography, echocardiography, ultrasound, conventional CT, UFCT, MRI, MRIS, PET, SPECT, other types of medical data, or combinations of the data, that are obtained from a particular medical monitoring device, could be inherently 1-, 2-, 3- or higher-dimensional data.
- the data is reformatted into a two-dimensional matrix.
- the data When data of three or higher dimensions is obtained, the data is separated into multiple two-dimensional matrices, which are then concatenated along a common dimension, forming one large twodimensional matrix.
- data that is inherently one-dimensional is obtained, it is first segmented and then reformatted into a two-dimensional matrix.
- ECG data obtained from a single sensor lead is inherently one-dimensional data, the single dimension being time.
- the data is reformatted by collecting it and segmenting it into individual heartbeats, which include the P-, QRS- and T-waves .
- the individual heartbeats are then arranged as rows of a matrix, resulting in two ⁇ dimensional data, one dimension being heartbeat waveform and the other time history.
- the three-dimensional data is separated into multiple two ⁇ dimensional matrices, one for each of the measured distances, for example each matrix having dimensions vertical angle by horizontal angle. These matrices are then concatenated together along a common dimension, such as horizontal angle, to form a concatenated two-dimensional matrix having horizontal angle in one dimension and a combination of vertical angle and distance in the other dimension.
- the three-dimensional data may be reformatted into two-dimensional form in other ways .
- the data are inherently 2-dimensional, generally having sagittal and transverse dimensions.
- M-scan or B-scan ultrasound data are collected, the data are also inherently 2- dimensional, generally having sagittal and temporal dimensions.
- MRIS data are collected, the data are also inherently two-dimensional, generally having the dimensions spectral frequency and time.
- phonocardiology the data are inherently two-dimensional, generally having dimensions time by frequency.
- multiple sensor phonocardiology the data are inherently three-dimensional, generally having dimensions time by frequency by location.
- CT, UFCT, MRI and SPECT data are collected, the data are inherently 3- dimensional, usually having the dimensions sagittal, transverse, and time.
- M-scan and B- scan ultrasound data are collected, the "fused" data are inherently 3-dimensional, having the dimensions sagittal, transverse and time.
- three-dimensional data such as 3D MRI data
- this data becomes four-dimensional (X, Y, Z and time)
- four- dimensional data is also reformatted into a single two ⁇ dimensional matrix.
- Other modalities may include higher- dimensional data, which are similarly reformatted.
- the matrix is weighted and concatenated with the history data representing features of interest f ⁇ , Rfi, weights, and model data R m (shown at 24) .
- History data is a special set of medically-pertinent features of interest in the form of right and/or left singular vectors that have been saved in a database.
- the history database is used when input signals are obtained for more than one time interval, planar slice, or lead. In such a case, the data obtained during each time interval, planar slice, or sensor lead is concatenated with the history data, and the history data is updated at selected intervals.
- the history data singular vectors represent features of interest Rtf£, f ⁇ ; features not of interest Rt fn , Lf n ; enhanced features of interest Rt efn» L efi ; an ⁇ ⁇ enhanced features not of interest Rt e f n ,
- the histories of features of interest Lfi are determined, in the preferred embodiment, through analysis of the singular values by the Subspace Selection Function.
- the features of interest may be determined by analysis of the singular vectors, or by some other criteria either in the Subspace Selection Function,
- history data of features of interest tfi, Lfi are calculated by the Subspace Selection
- Preprocessing Function 12 to be concatenated with the new medical data.
- the purpose of this feedback is to enable the energy from weak features of interest to build up to the point where the features of interest may be distinguished from background noise and to improve the description of features not of interest for removal from within the
- the history database may be saved and used at a later time to look for changes or differences between it and new medical data.
- the history database of a mammogram may be saved and compared six months later with a current mammogram to determine changes which may be indicative of cancer.
- the history data of features not of interest Rt n , I_f n are also calculated by the Subspace Selection Function 13 and are fed back to the Preprocessing Function 12 and forward to the Subspace Enhancement Function 14 and to the Signal Processing Function 15.
- enhanced features not of interest Rt e f n , I « e f n are fed back from the Subspace Enhancement Function 14 and/or the Signal Processing Function 15 to the Preprocessing Function 12.
- This information is used to set parameters for further analysis or enhancement and in alternative embodiments may be used to show dominant features in the data (e.g. respiration effects) without obscuring features of interest, or to reconstruct the original data.
- the history data is updated at selected time intervals by the Subspace Selection Function, the Subspace Enhancement Function and the Signal Processing Function.
- FIG. 3 a portion of the history database is shown.
- Right singular vectors representing features of interest Rtfi ⁇ OO 31, Rt fi.001 32 and Rt£ ⁇ Q2 33 for three time intervals right singular vectors representing features not of interest R f n .000 34 ' Rt fn.001 35 ' and Rt fn.002 36 for the same three time intervals, and right singular vector Rt m representing model features similarly stored in the history database 30.
- the history database 30 is updated for every time interval, for every planar slice, or every sensor lead, by storing the most recently received history data for right singular vector features of interest Rtfi.000 31.
- the pre-existing Rtfi QOO becomes Rtfi QOl 32 and the pre-existing Rtfi QOl becomes tfi_002 33, and so on ⁇ to Rt fi.nnn where "nnn" corresponds to the length of the compressed time history.
- the right singular vectors for features not of interest tf n _oOO 34, Rt fn.001 35, and Rt fn.002 6 are similarly stored and updated.
- the enhanced right singular vectors, left singular vectors, and enhanced left singular vectors are similarly stored in the history database.
- the history database thus contains historic data that shows the status of the input data for several previous time intervals, planar slices, or sensor leads.
- the history database is efficiently maintained in compressed and enhanced form, and represents features of the medical data at various time intervals, planar slices, or sensor leads.
- Each new analysis includes the compressed and enhanced historical data, which is equivalent to the benefits of a complete analysis of the full (uncompressed) historical data, yet at a fraction of the computational cost.
- the compressed history may be efficiently stored for later retrieval, review, analysis and comparison with new medical information.
- the compressed history database may be quantized, thresholded or otherwise modified for further compression to save storage space. Compression methods well known in the prior art, such as Run Length Coding (RLC) , Adaptive Differential Pulse Code Modification (ADPCM) , Huffman coding, or other compression techniques, may be used for further compression.
- RLC Run Length Coding
- ADPCM Adaptive Differential Pulse Code Modification
- Huffman coding Huffman coding
- the compressed history may be efficiently stored, for example in a "smart card” format for convenience.
- a "smart card” is a small, portable device such as the size of a credit card, which stores information in volatile RAM, in permanent ROM, or in laser data CDROM format.
- Such a “smart card” could then be retained in a patient's file or retained and carried by the patient in a credit card or health insurance card format.
- the compressed medical data may be retained in a central repository for later retrieval, transmission, expansion and analysis.
- model data is a special set of medically pertinent features of interest in the form of left and/or right singular vectors Rt m which have been predetermined by singular value decomposition of model-based data.
- a priori medical information may be used to guide the weight coefficients of the data to accentuate features of interest.
- singular value decomposition of the new medical data will describe features of interest within the data itself, by concatenating model-based singular vectors, correlations between the new medical data and the model- based data will enhance pertinent features of interest within the new medical data which are correlated with the model-based data.
- the resulting left singular vectors associated with the model-based information j n will indicate the degree of correlation between the model (s) and the new medical data.
- alternate models representing different medical hypotheses or potential diagnoses .
- the associated coefficients can then be used to evaluate the relevance of the various models to the observed new medical data.
- the Subspace Selection Function 13 is shown in greater detail.
- the two-dimensional data matrix X containing the concatenated input data, the history data and the model data, is obtained from the Preprocessing Function (shown at 41) .
- the two-dimensional data matrix X is decomposed using singular value decomposition at 42 to obtain its corresponding singular vectors , Rt and singular values D.
- the singular values and singular vectors are used to efficiently describe patterns found in the data.
- Matrix analysis using singular values and singular vectors is well known in the prior art.
- the following publications describe such matrix analysis in detail: Digital Spectral Analysis with Applications, S.L. Marple, 1987; Matrix Computations, G.H. Golub and C.F. Van Loan, 1989; "Singular Value Decomposition and Least Squares Solutions,” Numerical Math, G.H. Golub and C. Reisch, 1970; LINPAC User's Guide, J.J. Dongerra, et. al. , 1979; and "An Improved Algorithm for Computing Singular Value Decomposition," T.F. Chan, Communications of the ACM, 1982.
- the left and right singular vectors or sets of singular vectors represent specific characteristics of the data, and, by modifying certain of the singular values and/or singular vectors, selected features within the raw data may be isolated, enhanced, or suppressed.
- the left singular vectors L correspond to the vertical axis of the data matrix X
- the right singular vectors R correspond to the horizontal axis of the data matrix.
- the singular values D describe the magnitude of the associated features.
- the left singular vectors L describe features in the ECG data in terms of their temporal history
- the right singular vectors R- describe features in terms of their waveform structure.
- electrical alternans a feature of interest in ECG analysis of ischemia and tachycardia, show up by examining differences between heartbeats over time, they can be detected by analyzing the left singular vectors L.
- Ventricular late potentials, an indicator of potential tachycardia, however, are detected by examining the waveform structure, represented by the right singular vectors R.
- the left singular vectors describe features in the data in terms of location or cranial position, while the right singular vectors describe the temporal history and waveform structure, which may be used for analysis of abnormal electrical patterns .
- the right singular vectors describe features in the data in terms of their frequency composition, while the left singular vectors describe the temporal history of the features .
- the right singular vectors describe features in the data in terms of their frequency composition, while the left singular vectors describe a combination of temporal history and location of the features .
- the SVD data compression generally results in vectors that describe essential characteristics of the data in the sagittal and transverse dimensions.
- M-scan ultrasound data is analyzed
- the SVD data compression represents features of the data in terms of the spectral and temporal dimensions.
- UFCT, MRI and SPECT data with cine-loop processing
- the SVD data compression represents features of the data in terms of their sagittal, transverse, and temporal dimensions.
- the singular values D describe the magnitude of the associated features.
- left singular vectors L, right singular vectors Rt, and singular values D are thus used to represent important features within the input data, but in a substantially compressed form.
- This data compression allows the data to be enhanced, further processed, and displayed, without losing any necessary data, but saving substantial amounts of time and computing resources .
- SVD has an added advantage. Because of the orthonormality property of singular vectors, the computationally intensive process of matrix inversion in CT projections such as conventional CT, SPECT and UFCT, may be achieved through a simple matrix multiplication by the transposed singular vectors, thereby greatly reducing the computations involved in tomographic image projection.
- R-l Q r D r ⁇ l pt. Because D is a diagonal matrix, D ⁇ l is easily determined by taking the reciprocal of its diagonal elements .
- the data processing system uses singular value decomposition to describe patterns, remove unwanted components, and isolate and analyze components of interest in medical data.
- eigenvector decomposition of the cross-product matrix of the medical data may be used to decompose the medical data.
- Eigenvector decomposition is also well known in the prior art.
- the singular values are classified into subspaces at 43 based on the magnitudes of the singular values or by some other criteria.
- singular vectors are classified as one of three general categories: features not of interest (e.g. dominant, yet medically unimportant characteristics such as respiration effects, and effects of AC interference); (2) medical features of interest; and (3) noise.
- features not of interest e.g. dominant, yet medically unimportant characteristics such as respiration effects, and effects of AC interference
- medical features of interest e.g. dominant, yet medically unimportant characteristics such as respiration effects, and effects of AC interference
- noise There may be different subspace categories in alternative embodiments.
- preliminary classification may occur in the Subspace Filtering Function, while further, more sophisticated classification may occur in the Signal Processing Function.
- the Subspace Selection Function classifies the singular vectors into subspaces through various means, including:
- the data is enhanced at 44 by modifying singular values according to the subspace categories. For example, singular values corresponding to the noise subspace and features not of interest are reduced or set to zero, while the singular values corresponding to the subspace of features of interest remain unchanged or are increased.
- the data is analyzed in terms of its dominant singular values D and associated left and right singular vectors , R. The adjustment of the singular values assists in noise reduction, reduction of artifacts and other features not of interest, as well as contrast enhancement of features of interest.
- singular vectors are rotated using varimax rotation, quartimax rotation, oblique rotation, or other criteria, to maximize their loadings to accentuate data at particular locations in the matrix corresponding to different structures within the body or signals from the body. These operations are well known in the prior art.
- the Subspace Selection Function 13 modifies, refines the data, and selects subsets of singular vectors.
- a selected set of singular vector history data is passed on to the Subspace Enhancement Function for further processing and analysis of pertinent medical information contained in the singular vectors.
- the selected set is also saved in the history database to be passed back to the Signal Processing Function for processing and analysis of data in subsequent analyses.
- Fig. 5a is a flow chart showing an alternate embodiment of the Subspace Enhancement Function 14 in greater detail.
- the Subspace Enhancement Function uses analysis of the autocorrelation function of the singular vectors to reduce the effects of noise and other artifacts and to enhance the contrast of features of interest.
- the selected compressed and refined data in the form of singular vectors L, R fc , are passed by the Subspace Selection Function 13 to the Subspace Enhancement Function 14 for further enhancement of medically important indicators within the data.
- the singular vectors are converted at 51 into their Hankel matrix form.
- a Hankel matrix is constructed by placing multiple copies of the selected vector in a matrix, with each copy offset to the left one sample, relative to the other copies. After each singular vector is converted to a Hankel matrix, the Hankel matrices are concatenated together at 52 into a single matrix H.
- Fig. 6 shows a singular vector 68.
- Fig. 7 shows the singular vector 68 transformed into its Hankel matrix form 69, and
- Fig. 8 shows three Hankel matrices 81, 82, 83 concatenated together to form a single Hankel matrix H 84.
- the matrix B is decomposed using singular value decomposition at 53 into singular vectors P, Q and singular values E, although eigenvector decomposition (EVD) of the products HtH or HHt could be used interchangeably.
- ELD eigenvector decomposition
- the Hankel matrix singular vectors P, Q and singular values E are partitioned into subspaces at 53 to eliminate undesirable features, and to further enhance features of interest. Subspace selection may be accomplished through analysis of singular values, analysis of the singular vectors, or other criteria.
- Enhanced singular vectors L e fj , L e f n are generated at 54 by combining together selected singular vectors P, which have been weighted by their corresponding singular values E or other criteria:
- the autocorrelation function of each singular vector is calculated using the Fast Fourier Transform (FFT) at 61.
- FFT Fast Fourier Transform
- IFFT Inverse FFT
- the resulting autocorrelation function is then placed into its Toeplitz matrix form A.
- the autocorrelation matrix A is decomposed using singular value decomposition at 62.
- the decomposition describes the autocorrelation structures in terms of singular values E and singular vectors P.
- Selected singular vectors P s may be used as FIR filter coefficients to filter the original vector (e.g. Rt) , creating a filtered singular vector Rft.
- Rt original vector
- Rft filtered singular vector
- a small set of singular vectors describes the dominant structural characteristics of Rt.
- the remaining singular vectors are often related to noise, and thus may be modified, reduced or eliminated.
- Enhanced singular vectors L e fi. L e f n are generated at 64 by combining together selected singular vectors P, which have been weighted by their corresponding singular values E or other criteria:
- Fig. 9a is a graphic display of a singular vector
- Fig. 9b is a graphic display of Ps
- the FIR filter coefficients is the FIR filter coefficients
- Fig. 9c is the Subspace-Enhanced form of the singular vector g fi 94 of Fig. 9a.
- Enhanced singular vectors Rt e n are generated by combining together selected singular vectors Qt, which have also been weighted by their corresponding singular values E:
- Fig. 10 shows the Signal Processing Function 15 in greater detail.
- the compressed and enhanced data in the form of enhanced singular vectors, L e and R fc e are passed from the Subspace Enhancement Function 14 for further analysis.
- the singular vectors are subjected to advanced signal processing at 101.
- the advanced signal processing could be spectral processing (e.g. for ECG data), spectral-temporal processing (e.g. for phonocardiographic data or MRIS data) , or inverse solution (e.g. for CT image projection and reconstruction) .
- the singular vectors are convolved for "deblurring" purposes, then subsequently expanded through the technique of back projection.
- advanced signal processing techniques operate on compressed and enhanced singular vectors L e , R fc e rather than on raw data.
- L e singular vectors
- R fc e singular vectors
- Fourier techniques may be used to identify electrical alternans through analysis of L e and/or ventricular late potentials through analysis of R e fc .
- phonocardiology Fourier techniques may be used to identify abnormalities in the function of heart valves.
- matched filter processing of singular vectors and modification of singular values may be used to enhance the form and contrast resolution of weak, yet medically important features.
- the singular vectors may be convolved for the purpose of "deblurring” , then processed further to generate back projections of internal structures.
- the input to the Signal Processing Function is in the form of enhanced singular vectors L e , Rt e , rather than raw input data.
- Fig. 11 shows the Subspace Expansion Function 16 in greater detail.
- the enhanced data X e fi, x efn displayed by the Display Function 17.
- the data which has been enhanced and/or thresholded, is displayed so that features that occur in the data are displayed in the format in which they appear inherently in the data.
- the two-dimensional concatenated matrix is reformatted back into its original form.
- Three-dimensional data is displayed as a threedimensional data cube.
- Four-dimensional data is displayed using color in the three-dimensional cube.
- features of interest are displayed in the form of opaque objects within an otherwise transparent cube on an operator interface screen.
- Features not of interest may similarly be displayed in a separate display.
- the operator can interactively rotate the three-dimensional transparent cube through various angles and display the data from different perspectives. This type of display is not possible in prior art systems which do not eliminate interference and noise, nor correlate enhanced data across all dimensions .
- the singular vector SV X would first be multiplied by the appropriate scalar value SVy a( -j 138, then expanded into its full two-dimensional form by postmultiplying by the transpose vector S "t za . with the prior rotation of singular vectors, considerable computations may be saved, as contrasted to the rotation of the full, three-dimensional dataset.
- the screen display includes cursors for each of the dimensions of the data, which together allow a diagnostician to freely “travel” through the cube, displaying any desired data at any time.
- the diagnostician can use these cursors to control the display of "slices, " or planes, through the data.
- the diagnostician can display these slices, or planes, in the three-dimensional data cube using a set of multiple planar displays.
- the "slices" may present the enhanced data without thresholding, thereby displaying enhanced features against an otherwise opaque background.
- additional displays are used to illustrate individual singular vectors, or sets of singular vectors.
- Other displays are used to illustrate the frequency spectra of the singular vectors, or in the case of the right singular vectors, spectrogram displays.
- a "movie" time history of data received for each time interval is displayed by thresholding and rapidly expanding the compressed data into its full, three-dimensional form and displaying the data in a time sequence within an evolving, transparent data cube. Furthermore, as new data is added to the top of the data cube, older time history data is dropped off the bottom of the data cube.
- a movie display of enhanced data against an opaque background may be displayed as if one were rapidly "slicing" through the data cube.
Abstract
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP94915982A EP0699048A4 (en) | 1993-05-06 | 1994-05-03 | Method and system to enhance medical signals |
JP6525504A JPH09500990A (en) | 1993-05-06 | 1994-05-03 | Method and system for enhancing medical signals |
AU67805/94A AU674958B2 (en) | 1993-05-06 | 1994-05-03 | Method and system to enhance medical signals for real-time analysis and high-resolution display |
KR1019950704902A KR960702274A (en) | 1993-05-06 | 1995-11-06 | METHOD AND SYSTEM TO ENHANCE MEDICAL SIGNALS |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US08/058,571 | 1993-05-06 | ||
US08/058,571 US5490516A (en) | 1990-12-14 | 1993-05-06 | Method and system to enhance medical signals for real-time analysis and high-resolution display |
Publications (1)
Publication Number | Publication Date |
---|---|
WO1994026163A1 true WO1994026163A1 (en) | 1994-11-24 |
Family
ID=22017655
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US1994/004846 WO1994026163A1 (en) | 1993-05-06 | 1994-05-03 | Method and system to enhance medical signals |
Country Status (8)
Country | Link |
---|---|
US (1) | US5490516A (en) |
EP (1) | EP0699048A4 (en) |
JP (1) | JPH09500990A (en) |
KR (1) | KR960702274A (en) |
AU (1) | AU674958B2 (en) |
CA (1) | CA2161976A1 (en) |
IL (1) | IL109566A (en) |
WO (1) | WO1994026163A1 (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0759598A2 (en) * | 1995-08-21 | 1997-02-26 | AT&T Corp. | Image processing system and method |
US5690118A (en) * | 1995-09-01 | 1997-11-25 | Siemens Elema Ab | Method and apparatus for correcting non-physiological variations in ECG signals |
CN116525073A (en) * | 2023-07-03 | 2023-08-01 | 山东第一医科大学第一附属医院(山东省千佛山医院) | Database intelligent management system based on health physical examination big data |
Families Citing this family (77)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6113587A (en) * | 1990-09-24 | 2000-09-05 | Plc Medical Systems, Inc. | Handpiece for a medical laser system |
US5641965A (en) * | 1992-11-20 | 1997-06-24 | British Technology Group Limited | Image reconstruction |
EP0638869B1 (en) * | 1993-08-13 | 1995-06-07 | Siemens Aktiengesellschaft | Procedure for a high-resolution spectral analysis for multichannel observations |
US5757664A (en) * | 1996-06-04 | 1998-05-26 | Warren Rogers Associates, Inc. | Method and apparatus for monitoring operational performance of fluid storage systems |
US6925397B2 (en) | 1994-11-29 | 2005-08-02 | Warren Rogers Associates | Meter calibration accuracy |
US5887588A (en) * | 1995-02-23 | 1999-03-30 | Usenius; Jussi-Pekka R. | Automated method for classification and quantification of human brain metabolism |
US6691061B1 (en) | 1996-06-04 | 2004-02-10 | Warren Rogers Associates, Inc. | Method and apparatus for monitoring operational performance of fluid storage systems |
US6934644B2 (en) * | 1996-06-04 | 2005-08-23 | Warren Rogers Associates, Inc. | Method and apparatus for monitoring operational performance of fluid storage systems |
US6909986B2 (en) * | 1996-06-04 | 2005-06-21 | Warren Rogers Associates, Inc. | Method and apparatus for monitoring operational performance of fluid storage systems |
US6993186B1 (en) * | 1997-12-29 | 2006-01-31 | Glickman Jeff B | Energy minimization for classification, pattern recognition, sensor fusion, data compression, network reconstruction and signal processing |
DE19821761A1 (en) * | 1998-05-14 | 1999-11-18 | Braun Melsungen Ag | Method and device for the compressed optical representation of medical data |
US6152883A (en) * | 1998-06-23 | 2000-11-28 | Dalhousie University | KLT-based quality controlled compression of a single lead egg |
US6370416B1 (en) * | 1998-11-25 | 2002-04-09 | Ge Medical Systems Global Technology Company Llc | fMRI signal processing |
US6393090B1 (en) | 1998-12-31 | 2002-05-21 | General Electric Company | Computed tomography scout images with depth information |
WO2000052604A1 (en) * | 1999-03-05 | 2000-09-08 | Stayhealty. Com | System and method for on-line health monitoring and education |
US8103325B2 (en) * | 1999-03-08 | 2012-01-24 | Tyco Healthcare Group Lp | Method and circuit for storing and providing historical physiological data |
DE60041577D1 (en) * | 1999-03-08 | 2009-04-02 | Nellcor Puritan Bennett Llc | PROCESS AND CIRCUIT FOR STORAGE AND READY |
US20020009394A1 (en) | 1999-04-02 | 2002-01-24 | Hubert Koster | Automated process line |
FR2793055B1 (en) * | 1999-04-29 | 2001-07-13 | Ge Medical Syst Sa | METHOD AND SYSTEM FOR MERGING TWO DIGITAL RADIOGRAPHIC IMAGES |
CA2374040C (en) * | 1999-07-02 | 2010-10-19 | Hypermed Imaging, Inc. | Integrated imaging apparatus |
US6984207B1 (en) * | 1999-09-14 | 2006-01-10 | Hoana Medical, Inc. | Passive physiological monitoring (P2M) system |
US6440082B1 (en) * | 1999-09-30 | 2002-08-27 | Medtronic Physio-Control Manufacturing Corp. | Method and apparatus for using heart sounds to determine the presence of a pulse |
US9248306B2 (en) | 1999-09-30 | 2016-02-02 | Physio-Control, Inc. | Pulse detection apparatus, software, and methods using patient physiological signals |
US20040039419A1 (en) * | 1999-09-30 | 2004-02-26 | Stickney Ronald E. | Apparatus, software, and methods for cardiac pulse detection using a piezoelectric sensor |
US20030109790A1 (en) * | 2001-12-06 | 2003-06-12 | Medtronic Physio-Control Manufacturing Corp. | Pulse detection method and apparatus using patient impedance |
US7917301B1 (en) | 2000-09-19 | 2011-03-29 | Sequenom, Inc. | Method and device for identifying a biological sample |
US6421552B1 (en) * | 1999-12-27 | 2002-07-16 | Ge Medical Systems Global Technology Company, Llc | Methods and apparatus for estimating cardiac motion using projection data |
US7666151B2 (en) | 2002-11-20 | 2010-02-23 | Hoana Medical, Inc. | Devices and methods for passive patient monitoring |
US6813373B1 (en) * | 2001-04-03 | 2004-11-02 | Koninklijke Philips Electronics, N.V. | Image segmentation of embedded shapes using constrained morphing |
US7356188B2 (en) * | 2001-04-24 | 2008-04-08 | Microsoft Corporation | Recognizer of text-based work |
US7020775B2 (en) | 2001-04-24 | 2006-03-28 | Microsoft Corporation | Derivation and quantization of robust non-local characteristics for blind watermarking |
US6973574B2 (en) * | 2001-04-24 | 2005-12-06 | Microsoft Corp. | Recognizer of audio-content in digital signals |
DE10121802B4 (en) * | 2001-05-04 | 2004-09-30 | Siemens Ag | Process for image generation in a magnetic resonance imaging device and magnetic resonance imaging device and data processing system |
ITSV20010017A1 (en) | 2001-05-28 | 2002-11-28 | Esaote Spa | DEVICE FOR DETECTION OF IMAGES IN NUCLEAR MAGNETIC RESONANCE |
US6675036B2 (en) | 2001-07-18 | 2004-01-06 | Ge Medical Systems, Inc. | Diagnostic device including a method and apparatus for bio-potential noise cancellation utilizing the patient's respiratory signal |
WO2003049032A2 (en) * | 2001-12-07 | 2003-06-12 | Koninklijke Philips Electronics N.V. | Medical viewing system and method for spatially enhancing structures in noisy images |
CA2436400A1 (en) * | 2002-07-30 | 2004-01-30 | Abel G. Wolman | Geometrization for pattern recognition, data analysis, data merging, and multiple criteria decision making |
US20040039420A1 (en) * | 2002-08-26 | 2004-02-26 | Medtronic Physio-Control Manufacturing Corp. | Apparatus, software, and methods for cardiac pulse detection using accelerometer data |
US20040116969A1 (en) | 2002-08-26 | 2004-06-17 | Owen James M. | Pulse detection using patient physiological signals |
US7429860B2 (en) * | 2003-01-28 | 2008-09-30 | University Of Southern California | Noise reduction for spectroscopic signal processing |
DE10328765B4 (en) * | 2003-06-25 | 2005-11-24 | aviCOM Gesellschaft für angewandte visuelle Systeme mbH | Device and method for connecting the representation of the electric cardiac field with the representation of the associated heart |
JP4506119B2 (en) * | 2003-07-09 | 2010-07-21 | 株式会社島津製作所 | Digital subtraction apparatus, digital angiography apparatus, and digital subtraction image display method |
US7300405B2 (en) * | 2003-10-22 | 2007-11-27 | 3M Innovative Properties Company | Analysis of auscultatory sounds using single value decomposition |
US20050119866A1 (en) * | 2003-11-14 | 2005-06-02 | Zaleski John R. | Medical parameter processing system |
US20050111757A1 (en) * | 2003-11-26 | 2005-05-26 | Brackett Charles C. | Auto-image alignment system and method based on identified anomalies |
US7831832B2 (en) * | 2004-01-06 | 2010-11-09 | Microsoft Corporation | Digital goods representation based upon matrix invariances |
US20050165690A1 (en) * | 2004-01-23 | 2005-07-28 | Microsoft Corporation | Watermarking via quantization of rational statistics of regions |
US7574333B2 (en) * | 2004-02-05 | 2009-08-11 | Honeywell International Inc. | Apparatus and method for modeling relationships between signals |
US7363200B2 (en) * | 2004-02-05 | 2008-04-22 | Honeywell International Inc. | Apparatus and method for isolating noise effects in a signal |
ATE374979T1 (en) * | 2004-04-13 | 2007-10-15 | Aic | FRACTAL SKR METHOD FOR EVALUATION OF IMAGE QUALITY |
US7770014B2 (en) * | 2004-04-30 | 2010-08-03 | Microsoft Corporation | Randomized signal transforms and their applications |
FR2872919B1 (en) * | 2004-07-09 | 2006-09-29 | Thales Sa | FAST COHERENT TREATMENT FOR PERIODIC RATE SPECTRUM CODES |
US7288066B2 (en) * | 2004-11-01 | 2007-10-30 | Medtronic, Inc. | Data compression method for implantable medical devices |
US20060167385A1 (en) * | 2005-01-24 | 2006-07-27 | 3M Innovative Properties Company | Analysis of auscultatory sounds using voice recognition |
KR100772373B1 (en) * | 2005-02-07 | 2007-11-01 | 삼성전자주식회사 | Apparatus for data processing using a plurality of data processing apparatus, method thereof, and recoding medium storing a program for implementing the method |
US8038625B2 (en) * | 2005-09-15 | 2011-10-18 | St. Jude Medical, Atrial Fibrillation Division, Inc. | System and method for three-dimensional mapping of electrophysiology information |
US8229545B2 (en) | 2005-09-15 | 2012-07-24 | St. Jude Medical, Atrial Fibrillation Division, Inc. | System and method for mapping complex fractionated electrogram information |
US20070076869A1 (en) * | 2005-10-03 | 2007-04-05 | Microsoft Corporation | Digital goods representation based upon matrix invariants using non-negative matrix factorizations |
US7988639B2 (en) * | 2006-05-17 | 2011-08-02 | St. Jude Medical, Atrial Fibrillation Division, Inc. | System and method for complex geometry modeling of anatomy using multiple surface models |
US7860304B2 (en) * | 2006-12-11 | 2010-12-28 | Canon Kabushiki Kaisha | Constructing basis functions using sensor wavelength dependence |
US20080167638A1 (en) * | 2007-01-04 | 2008-07-10 | International Business Machines Corporation | Method and Apparatus for Delivering Pharmaceuticals Based on Real-Time Monitoring of Tissue States |
US8306296B2 (en) * | 2009-04-30 | 2012-11-06 | Medison Co., Ltd. | Clutter signal filtering using eigenvectors in an ultrasound system |
US20110224914A1 (en) * | 2010-03-10 | 2011-09-15 | Sharp Laboratories Of America, Inc. | Constrained processing technique for an impedance biosensor |
WO2012098844A1 (en) | 2011-01-18 | 2012-07-26 | パナソニック株式会社 | Vehicle-direction identification device, vehicle-direction identification method, and program therefor |
DE102013205278A1 (en) * | 2013-03-26 | 2014-10-02 | Siemens Aktiengesellschaft | Method for displaying signal values of a combined magnetic resonance positron emission tomography device and correspondingly designed magnetic resonance positron emission tomography device |
US9812150B2 (en) * | 2013-08-28 | 2017-11-07 | Accusonus, Inc. | Methods and systems for improved signal decomposition |
DE102014203057B4 (en) * | 2014-02-20 | 2018-12-27 | Siemens Healthcare Gmbh | Method for determining a radio-frequency pulse by means of vector representation of sample points |
US20150264505A1 (en) | 2014-03-13 | 2015-09-17 | Accusonus S.A. | Wireless exchange of data between devices in live events |
US10468036B2 (en) | 2014-04-30 | 2019-11-05 | Accusonus, Inc. | Methods and systems for processing and mixing signals using signal decomposition |
CN107843862B (en) * | 2016-09-20 | 2021-04-23 | 奥泰医疗系统有限责任公司 | Non-iterative generation method of reference position image in PROPELLER technology |
KR101942854B1 (en) * | 2017-05-29 | 2019-04-17 | 성균관대학교산학협력단 | Device and method for magnet resonance imaging |
CN107174236B (en) * | 2017-06-21 | 2020-03-10 | 广东工业大学 | Electrocardiosignal denoising method and device based on optimization theory |
CN112347813A (en) * | 2019-08-07 | 2021-02-09 | 顺丰科技有限公司 | Baseline detection method, equipment and storage medium for high signal-to-noise ratio time series |
CN113180636B (en) * | 2021-04-29 | 2022-09-16 | 杭州微影医疗科技有限公司 | Interference cancellation method, medium, and apparatus |
CN113706425A (en) * | 2021-06-02 | 2021-11-26 | 上海深至信息科技有限公司 | Image enhancement system and portable Doppler imager |
WO2023064894A1 (en) * | 2021-10-15 | 2023-04-20 | Battelle Energy Alliance, Llc | Detection of behavior anomalies in control systems and related systems, methods, and devices |
CN113973090B (en) * | 2021-10-18 | 2023-12-08 | 北谷电子股份有限公司 | Apparatus and method for processing big data in communication network |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4422459A (en) * | 1980-11-18 | 1983-12-27 | University Patents, Inc. | Electrocardiographic means and method for detecting potential ventricular tachycardia |
Family Cites Families (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US422459A (en) * | 1890-03-04 | Whiffletree-coupling | ||
US3894348A (en) * | 1966-12-19 | 1975-07-15 | Us Navy | Relator simulator system |
US4458691A (en) * | 1982-02-11 | 1984-07-10 | Arrhythmia Research Technology, Inc. | System and method for predicting ventricular tachycardia by adaptive high pass filter |
US4458692A (en) * | 1982-02-11 | 1984-07-10 | Arrhythmia Research Technology, Inc. | System and method for predicting ventricular tachycardia with a gain controlled high pass filter |
US4492235A (en) * | 1982-02-11 | 1985-01-08 | Arrhythmia Research Technology, Inc. | System and method for predicting ventricular tachycardia by derivative analysis |
US4630204A (en) * | 1984-02-21 | 1986-12-16 | Mortara Instrument Inc. | High resolution ECG waveform processor |
GB2165727B (en) * | 1984-10-10 | 1988-05-18 | Plessey Co Plc | A method of selectively displaying data |
US4965732A (en) * | 1985-11-06 | 1990-10-23 | The Board Of Trustees Of The Leland Stanford Junior University | Methods and arrangements for signal reception and parameter estimation |
US5010504A (en) * | 1988-02-22 | 1991-04-23 | Eastman Kodak Company | Digital image noise suppression method using SVD block transform |
US4973111A (en) * | 1988-09-14 | 1990-11-27 | Case Western Reserve University | Parametric image reconstruction using a high-resolution, high signal-to-noise technique |
US4839853A (en) * | 1988-09-15 | 1989-06-13 | Bell Communications Research, Inc. | Computer information retrieval using latent semantic structure |
US5031155A (en) * | 1989-04-28 | 1991-07-09 | Schlumberger Technology Corporation | Compression and reconstruction of sonic data |
US4995011A (en) * | 1989-09-20 | 1991-02-19 | Woods Hole Oceanographic Institute | Acoustic mapping system using tomographic reconstruction |
US5109863A (en) * | 1989-10-26 | 1992-05-05 | Rutgers, The State University Of New Jersey | Noninvasive diagnostic system for coronary artery disease |
US5001747A (en) * | 1990-01-22 | 1991-03-19 | Racal Data Communications Inc. | Frequency offset estimator for data modem |
US5175710A (en) * | 1990-12-14 | 1992-12-29 | Hutson William H | Multi-dimensional data processing and display |
US5348020A (en) * | 1990-12-14 | 1994-09-20 | Hutson William H | Method and system for near real-time analysis and display of electrocardiographic signals |
US5245587A (en) * | 1990-12-14 | 1993-09-14 | Hutson William H | Multi-dimensional signal processing and display |
-
1993
- 1993-05-06 US US08/058,571 patent/US5490516A/en not_active Expired - Fee Related
-
1994
- 1994-05-03 JP JP6525504A patent/JPH09500990A/en active Pending
- 1994-05-03 WO PCT/US1994/004846 patent/WO1994026163A1/en not_active Application Discontinuation
- 1994-05-03 CA CA002161976A patent/CA2161976A1/en not_active Abandoned
- 1994-05-03 AU AU67805/94A patent/AU674958B2/en not_active Ceased
- 1994-05-03 EP EP94915982A patent/EP0699048A4/en not_active Ceased
- 1994-05-05 IL IL10956694A patent/IL109566A/en not_active IP Right Cessation
-
1995
- 1995-11-06 KR KR1019950704902A patent/KR960702274A/en not_active Application Discontinuation
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4422459A (en) * | 1980-11-18 | 1983-12-27 | University Patents, Inc. | Electrocardiographic means and method for detecting potential ventricular tachycardia |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0759598A2 (en) * | 1995-08-21 | 1997-02-26 | AT&T Corp. | Image processing system and method |
EP0759598A3 (en) * | 1995-08-21 | 1997-09-10 | At & T Corp | Image processing system and method |
US5690118A (en) * | 1995-09-01 | 1997-11-25 | Siemens Elema Ab | Method and apparatus for correcting non-physiological variations in ECG signals |
CN116525073A (en) * | 2023-07-03 | 2023-08-01 | 山东第一医科大学第一附属医院(山东省千佛山医院) | Database intelligent management system based on health physical examination big data |
CN116525073B (en) * | 2023-07-03 | 2023-09-15 | 山东第一医科大学第一附属医院(山东省千佛山医院) | Database intelligent management system based on health physical examination big data |
Also Published As
Publication number | Publication date |
---|---|
AU6780594A (en) | 1994-12-12 |
IL109566A0 (en) | 1994-08-26 |
IL109566A (en) | 1998-12-27 |
US5490516A (en) | 1996-02-13 |
JPH09500990A (en) | 1997-01-28 |
KR960702274A (en) | 1996-04-27 |
EP0699048A4 (en) | 1997-06-11 |
EP0699048A1 (en) | 1996-03-06 |
AU674958B2 (en) | 1997-01-16 |
CA2161976A1 (en) | 1994-11-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US5490516A (en) | Method and system to enhance medical signals for real-time analysis and high-resolution display | |
US5348020A (en) | Method and system for near real-time analysis and display of electrocardiographic signals | |
Unser et al. | A review of wavelets in biomedical applications | |
JP5014694B2 (en) | Method and apparatus for generating a multi-resolution framework for improving medical imaging workflow | |
Meyer-Bäse | Pattern Recognition and Signal Analysis in Medical Imaging | |
Heinlein et al. | Integrated wavelets for enhancement of microcalcifications in digital mammography | |
Lane et al. | Direct phase retrieval | |
Rohling et al. | A comparison of freehand three-dimensional ultrasound reconstruction techniques | |
Kiebel et al. | The general linear model | |
Laine | Wavelets in temporal and spatial processing of biomedical images | |
EP0952547A2 (en) | Vascular imaging with adaptive averaging | |
US7401006B2 (en) | Method and system for signal processing using a sparse approximation of the S-transform | |
US8275214B2 (en) | Image texture characterization of medical images | |
Gerig et al. | Medical imaging and computer vision: An integrated approach for diagnosis and planning | |
Zhang et al. | Clutter suppression in ultrasound: performance evaluation and review of low-rank and sparse matrix decomposition methods | |
US7355605B2 (en) | Method and system for automatic orientation of local visualization techniques for vessel structures | |
Csébfalvi | Prefiltered Gaussian reconstruction for high-quality rendering of volumetric data sampled on a body-centered cubic grid | |
Boucheham et al. | Piecewise linear correction of ECG baseline wander: a curve simplification approach | |
EP2218054A1 (en) | Process and apparatus | |
Unser | Wavelets, statistics, and biomedical applications | |
WO2022271601A1 (en) | Systems and methods for reverberation clutter artifacts suppression in ultrasound imaging | |
US7319788B2 (en) | Visualization of S transform data using principal-component analysis | |
US6307911B1 (en) | Fast hierarchical backprojection for 3D Radon transform | |
Udrea et al. | Nonlinear deterministic methods for computer aided diagnosis in case of kidney diseases | |
Pan | Consistency conditions and linear reconstruction methods in diffraction tomography |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AU BY CA JP KR KZ RU |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE |
|
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2161976 Country of ref document: CA |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1994915982 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 1994915982 Country of ref document: EP |
|
WWR | Wipo information: refused in national office |
Ref document number: 1994915982 Country of ref document: EP |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 1994915982 Country of ref document: EP |