Resumen |
In this work, a new computational method for lossless image compression is developed. The wavelet transform is used for setting automatically lifting filter coefficients at each decomposition level. Our proposal is based on the analysis of spectral characteristics of each image decomposition level, where a 1-NN classifier is used for optimizing the lifting filter coefficients in terms of the minimal compressed image entropy. In our experiments we compare the proposed method against standard wavelet filters CDF(2,2), CDF(2,4), CDF(4,2) and CDF(4,4) and linear prediction based filters such as LPC(4,2) and LPC(4,4). Experimental results show a good performance of the proposed method according to the transformed image entropy and the compressed image bitstream. |