L_∞ infinity constrained high fidelity compression algorithm for gray-level finger- print image

L 2 based image coders can not offer a tight bound on the maximum quantization error of a pixel. This may result in the loss of important details in finger print images. This paper presents a high fidelity algorithm with a given tight L ∞ bound for compressing gray level finger print images. A local direction based predictor is proposed considering the characteristics of finger print image. The adaptive adjustment of entropy coders under near lossless mode is also proposed. This idea may help DPCM type predictive near lossless image coders. Experiments with real finger print images show that by incorporating the proposed techniques into the near lossless version of CALIC, which is considered by many as a state of the art algorithm, we are able to increase its PSNR by 1.2dB and reduce its bit rate by 8 percent. With a tight ±3 bound, the proposed method obtained a 5∶1 compression ratio with obviously higher PSNR results than SPIHT, one of the best L 2 based wavelet coders.