Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUST928003 I4
Publication typeGrant
Publication dateNov 5, 1974
Filing dateJul 23, 1973
Priority dateMay 30, 1973
Publication numberUS T928003 I4, US T928003I4, US-I4-T928003, UST928003 I4, UST928003I4
InventorsR. Bahl
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Sequential learning predictor for an image compression system
US T928003 I4
Images(10)
Previous page
Next page
Description  (OCR text may contain errors)

DEFENSIVE PUBLICATION UNITED STATES PATENT OFFICE in the application as originally filed. The files or these applications are available to the public to: inspection and reproduction may be purchased for 30 cents a sheet.

Defensive Publication applications have not been examined as to the merits of alleged invention. The Patent 0 09 me! no assertion as to the novelty of the disclosed subject matter.

PUBLISHED NOVEMBER 5, 1974 Int. Cl. H04n 7/12 U.S. Cl. 235154 10 Sheets Drawing. 26 Pages Specification A system is disclosed for compacting digital image data by means of predictive coding. The prediction decision rule is developed by a heuristic sequential learning process from a sample image data set.

I! ll omoman SEOUENTML DATA Pnealcros an L A large set of two dimensional image points represented in binary form is processed in order to generate a prediction decision tree structure. The decision tree is stored within the predictor memory, then, each image point taken from a set of data that is to be compressed, is predicted in accordance with the stored prediction tree structure. The predicted value is then added, modulo-two, to the actual binary value of the image point. The resulting output of the modulo-two addition is then compressed by conventional means into data that is transmitted to a receiving unit. The receiving unit contains both a decoder for expanding the data, and an inverse predictor for creating a data pattern representative of the image that was transmitted.

L. R- BAHL ET AL Nov. 5, 1974 SEQUENTIAL LEARNING PREDICTOR FOR AN IMAGE COMPRESSION SYSTEM Original Filed lay 30, 1973 10 Sheets-Sheet 1 mokuammm 43:63am Eozwz ow E0085 N q m K ON 2 520 zo E55 $23 205222;; I. h fi 2 r1 "655m 55 MW mwooozm mm iizmnomw Eozwz .EzGEo :2 w QZEZS 2 2 w 2 0 n Nov. 5, 1974 L, R BAHL ETAL T928,003

SEQUENTIAL LERNING PREDICTOR FOR AN IMAGE COMPRESSION SYSTEM Original Filed lay 30, 973 10 Sheets-Sheet 3 FIG. 2

3 POINT PREDICTOR (PRIOR ART) --DOCUMENT 1* l 16 1? 1s 19 2o 21 22 i a5? e 9101123l 14 s 2 3 4 12 24 13 5 1 2' I 1 T' T k ++++++#+v 2% #:hffk 1w NOV. 5, BAHL ETAL.

SEQUENTIAL LEARNING PREDICTOR FOR AN IMAGE COMPRESSION SYSTEM Original Filed Ilay 30, 1973 10 Sheets-Sheet Nov. 5, 1974 R BAHL ETAL T928,003

SEQUENTIAL LEARNING PREDICTOR FOR AN IMAGE COMPRESSION SYSTEM Original Filed May 30, 1973 10 Sheets-Sheet 4 FIG. 7 Fl G 7A READ IN USER SPECIFIED H0 7A PARAMETERS AND DATA FIG.

SET TREE LEVEL INDEX T U? SET PREDICTTON TEST POINT INDEX =1 \WT APPLY CURRENT PREDICTION TEST TO ALL DATA AND STORE RESULTS SET NDDE INDEX AT CURRENT TREE LEVEL T0 i COMPUTE PERFDRMANCE OF CURRENT PREDICTIDN TEST FDR CURRENT NODE AND STORE RESULTS IS CURRENT TEST RESULT BETTER THAN BEST PREVIOUS PREDICTION TEST Nov. 5, 1974 Original Filed lay 30, 1973 L. R. BAHL. E'I'AL T928,003

SEQUENTIAL LEARNING PREDIC'IOR FOR AN IMAGE COMPRESSION SYSTEM 10 Sheets-Sheen INCREMENT NUDE NO INDEX BY I SUBSTITUTE CURRENT PREDICTION TEST FOR BEST PREVIOUS AT CURRENT NODE AND STORE ASSOCIATED PREDICTION DECISION IN DECISION TREE I26 fiLN NDDES EXAMINED HAVE INCREMENT PREDICTION l0 ALL POINTS TEST POINT INDEX BY I BEEN TESTED MAXIMUM TREE LEVEL REACHED INCREMENT TREE LEVEL INDEX BY I ELIMINATE NODES AT CURRENT LEVEL USING COST FUNCTION AND MODIFY DECISION TREE ACCORDINCLY DECREMENT TREE LEVEL INDEX BYI CURRENT TREE LEVEL INDEX =2 NOV. 5, 1914 R BAHL ETAL T928,003

SEQUENTIAL LEARNING PREDICTOR FOR AN IMAGE COMPRESSION SYSTEM Original Filed llay 30, 1973 10 Sheets-Sheet I:

FIG. L j ST ART Fl (5. 8 A

8A 4 "TEST 1 PROGRAM PROCESS FOR FIG. 2 IDIITESTI}ITEST GENERATING ITREE (SEQ 8B JDIITEST) =1,NTEST DEC TREE) TO BE USED 3 FOR PREDICTION NROW,NCOL IDATA(I,J),I-1,NROW J -i,NCOL

FIG. 8C

mma

FIG 8D Fl G. 8 III ITIALIZE ARRAY 2 2 IFLAG(,)-0

N1- HHSTO L-IDATAIILJJ) N2-2**(KSTOP1) 2 4 PIS-2* *(KSTOPH) ICOUNT I1F.I,IIII k I l -ICOUNT I1r,I,IIII+I I INITIALIZE ARRAY 208 ICOUNT ,I-o

2| 2 11- mmmsn 22s JJHHJDIITEST) ID =IDATMI ,J) IF =IFLAG(I ,J

YES

I I I N07. 5, 1974 BAHL' ETAL T928,003

SEQUENTIAL LEARNING PHEDICTOR FOR AN IMAGE COMPRIZSSION SYSTEM Original Filed lay 30, 1973 10 Sheets-Sheet FIG 9 1 NTEST PROGRAM FOR PREDICTION 2 mm JD (ITESU 3 NROW,NCOL 4 IDATA (I,J),I=4,NROW 400 J-4,HCOL

5 ITREE (lTADLITAD-LNS-i II=I+ID (ITEST) 422 JJ=J+JD(ITEST) FHA L=0 YES M NO L=IDATA(II,JJ)

YES ITAD- 2*ITAD+L DONE

Classifications
U.S. Classification709/247, 348/409.1
International ClassificationG06T9/00, H04N1/417
Cooperative ClassificationH04N1/417, G06T9/004
European ClassificationH04N1/417, G06T9/00P