Enhanced DPCM using LMS predictor and composite source model

A new algorithm combining an LMS predictor and composite source model is used to improve the performance of DPCM in the lossless compression of digital signals. The authors add a clipping function after the linear predictor to gain further improvement. Applications to image coding demonstrate its effectiveness and several useful conclusions are obtained.