| abstract |
THE PRESENT INVENTION CONCERNS METHODS OF COMPRESSION OF A DIGITIZED IMAGE. THE NUMERICAL VALUES B OF THE BRILLIANCES OF THE POINTS OF THE I LINE ARE SUBJECT TO A DIFFERENTIAL CODING WITH FOR REFERENCES THE VALUES OF BRILLIANCE B OR BI-1, J OF THE POINTS OF THE I-1 LINE. THEY ARE CODED BY BLOCKS OF CONSECUTIVE VALUES SUCH THAT THE DIFFERENCE IN BRIGHTNESS OF EACH POINT FROM ITS REFERENCE POINT IS IDENTICAL TO A TOLERANCE 2 NEAR. A BLOCK'S B VALUES ARE ALL ENCODED WITH RESPECT TO B VALUES OR ALL ENCODED WITH RESPECT TO B VALUES OR ALL ENCODED WITH RESPECT TO B VALUES, SO AS TO MINIMIZE THE DIFFERENCE IN BRIGHTNESS. EACH BLOCK IS CODE BY ITS BRILLIANCE DIFFERENCE VALUE, LENGTH, AND A BINARY WORD INDICTING IF THE RESPECTIVE REFERENCES ARE B OR BI-1, J-1 OR BI-1, J. APPLICATION, IN PARTICULAR, TO THE TRANSMISSION OF IMAGES AT REDUCED RATE AND TO THE STORAGE OF IMAGES ON VIDEO-DISCS. |