An efficient spatial prediction-based image compression scheme

Abstract

We have designed a spatial prediction-based image-compression scheme. The proposed scheme consists of two phases: the prediction phase and the quantization phase. In the prediction phase, a hierarchical structure among pixels in the image is built. Following the constructed hierarchical structure, the neighboring pixels are utilized to predict every central pixel. The prediction scheme generates an image map which indicates the prediction errors. The structure of the resulting image map is very similar to the result of a discrete wavelet transform. Thus, most quantization methods of wavelet or subband image-compression algorithms can be followed in our scheme directly to yield good compression performance. In the quantization phase, we design a multilevel threshold scheme to further enhance the result of SPIHT by taking the significance of the pixel values and the hierarchical levels into account. Furthermore, the proposed scheme can be realized by only a few integer additions and bit shifts. Simulation results indicate that the visual quality of the designed efficient spatial prediction-based image compression scheme is competitive with JPEG. All the above features make the designed image-compression scheme beneficial to the applications of real-time and wireless transmission in low-computational power environments.