Generally, image compression using Block Truncation Coding (BTC) can?t provide high compression ratio and visual quality. A new adaptive BTC scheme is proposed by combining visual perception of original images and data hiding technique. The texture sensitivity is exploited to classify the types of image sub-block and reduce the bit rate while remaining good quality of reconstructed images. Moreover, the indicator bits of image sub-blocks are hidden into the quantizing levels of those blocks for decoding purpose. The simulation results show that the proposed BTC algorithm can achieve good balance between bit-rate and PSNR values. Especially, it obtained about 4.0 dB higher PSNR at approximate bit-rate.
展开▼
