Lossless Image Compression by PPM-Based Prediction Coding

Lossless image compression, such as JPEG-LS and CALIC, are necessary for medical image and art work image archive, which has been processed by computers recently. Most of the conventional lossless image compressions are based on prediction compression which predicts the value of current symbol by using several symbols which have already been encoded. The difference between the predicted values and the actual values, called prediction error, is encoded by entropy coding independently. This paper proposes lossless image compression which is also based on prediction coding and the prediction errors are encoded by modified PPM (prediction by partial matching).