Compression of Computed Radiographic Images Using Linear Prediction on Wavelet Coefficients

Computed radiographic images are usually large in size. So for storage and transmission the image should be compressed. This paper proposes a method in which image compression is achieved by applying linear prediction on wavelet coefficients. The image is decomposed up to four levels using wavelet transform and on the detail coefficients prediction is performed. The compressed image comprises of predictor coefficients and retained coefficients. Performing linear prediction on them recovers the decompressed image. The scheme is evaluated using subjective and objective criteria and found to suit Computed Radiographic (CR) images.