Quadtree Image Compression Using Edge-Based Decomposition and Predictive Coding of Leaf Nodes

MCDNOP=QR Sf, 678ghij9kMDlmNn>S1deU:; ?^. ]SEF1\jJPEG0eU OPj27% jDoapN256×256 qCD0?rstNuSvwxSy^. ]SEF1\MJPEG5zMf,gh <1\0U{|{}SEFCD0U9~DN€‚ƒ„#U, b^VWXEFCDN4$ ^. AbstractThis paper proposes a quadtree image compression method which encodes images efficiently and also makes unartificial compressed images. The proposed compression method uses edge-based quadtree decomposition to preserve the significant edge-lines, and it utilizes the predictive coding scheme to exploit the high correlation of the leaf node blocks. The simulation results with 256×256 grayscale images verify that the proposed method yields better coding efficiency than the JPEG by about 25 percents. The proposed method can provide more natural compressed images as it is free from the ringing effect in the compressed images which used to be in the images compressed by the fixed block based encoders such as the JPEG.Keywords : Edge-Based Quadtree Decomposition; Image Compression