Lossless Compression of Medical Images Using a Dual Level DPCM with Context Adaptive Switching Neural Network Predictor

A novel dual level differential pulse code modulation (DL-DPCM) is proposed for lossless compression of medical images. The DL-DPCM consists of a linear DPCM followed by a nonlinear DPCM namely, context adaptive switching neural network predictor (CAS-NNP). The CAS-NNP adaptively switches between three NN predictors based on the context texture of the predicted pixel in the image. Experiments on magnetic resonance (MR) images showed lower prediction error for the DL-DPCM compared to the GAP and the MED, which are used in benchmark algorithms CALIC and LOCO-I respectively. The overall improvement in data reduction after entropy coding the prediction error were 0.21 bpp (6.5%) compared to the CALIC and 0.40 bpp (11.7%) compared to the LOCO-I.