A Hybrid Compression Algorithm for Compound Images

This paper presents a hybrid image coding scheme, based on shape primitives, termed Shape Primitive Extraction and Coding (SPEC). It is required that the compression algorithm should not only achieve high compression ratio, but also low complexity and high visual quality. The segmentation classifies image blocks into picture and text/graphics blocks by thresholding the number of colors of each block, and then extracts shape primitives of text/graphics from picture blocks. Dynamic color palette that tracks recent text/graphics colors is used to separate small shape primitives of text/graphics from pictorial blocks. Shape primitives are also extracted from text/graphics blocks. Then, the foreground text and graphics pixels is compressed with a lossless coding algorithm, which integrates shape-based and palette-based coding and the background pictorial pixels is compressed with lossy Joint Photographic Expert Group (JPEG). Experimental results demonstrate the efficiency and low complexity of the hybrid scheme. Keywords: Compression, SPEC, Compound Images