Bayer Pattern Images Compression Based on Fuzzy Neural Network Predictor
Gang Zhang · Journal of Taiyuan University of Technology · 2008
This paper has proposed a novel compression method based on raster-scan order focusing on the Bayer pattern images.The whole compression process has two phases,firstly,the Bayer pattern images are predicted by Takagi-Sugeno Fuzzy Neural Network(TFNN) directly,then the adaptive arithmetic coding is employed to encode the prediction errors,the average compression ratio can achieve 1.71∶1 above,the proposed compression method is of advantage to the existing methods,obviously.Meantime,it has characteristics of simple structure and easy reality by the hardware.