提升高精度图像弱目标视觉质量的高效压缩算法

为了实现对高精度图像进行高效压缩，同时确保重建图像的弱目标区域有较好的保真效果，提出了一种提升弱目标区域质量的基于误差优化编码的高精度图像压缩算法. 首先，使用JPEG-LS （joint photographic experts group lossless）压缩算法对图像数据进行压缩，在游程编码过程中自适应地选择需要二次编码的误差数据，并完成了基于视觉质量的非均匀量化；其次，对量化值进行数据分解，去除量化值之间的相关性，并对分解后的数据进行MQ算术编码的熵编码；图像重建时根据量化间隔重建反量化值，并设计了反量化优化和滤波优化过程；最后，将本文算法与JPEG-LS、JPEG2000 （joint photographic experts group 2000）算法进行了性能比较，结果表明:本算法能够实现高精度图像数据的高效压缩，且复杂度低，易于硬件实现，虽然引入了误差数据二次优化编码等过程，但增加编码的数据量较小，故与JPEG-LS算法的压缩速度相当，然而比JPEG2000算法的压缩速度提升4.47倍；同时有效减少了常规算法造成的信息损失，重建图像的峰值信噪比与JPEG-LS、JPEG2000相当或略低，但弱目标区域的视觉质量及保真效果更好. 

Efficient Compression Algorithm for Improving Visual Quality ofWeak Targets in High Precision Images

In order to effectively compress high precision images, and ensure desirable fidelity of weak target area of the reconstructed image, a high precision image compression algorithm based on error optimization coding is proposed to improve the quality of weak target area. First, the JPEG-LS (joint photographic experts group lossless) compression algorithm is used to compress the image data, and the error coding data is selected in an adaptive way during the run-length encoding process. Then, the non-uniform quantization based on visual quality is also carried out, and the quantized values are decomposed to remove the correlation between them. Finally, the entropy coding of the decomposed data is carried out by MQ arithmetic encoder. In the reconstruction, the inverse quantization value is reconstructed according to the quantized interval, and the inverse quantization optimization and filtering optimization are carried out. The performance of this algorithm is also compared with those of JPEG-LS and JPEG2000 (joint photographic experts group 2000) algorithms. The experimental results show that this algorithm can achieve high-efficiency and high-precision compression of image data. This algorithm is low in complexity and easy for hardware implementation. Although the algorithm incorporates error data optimization and coding process, the amount of data encoded is small, and it is equivalent to the JPEG-LS in compression speed. The compression speed of the JPEG2000 algorithm is about 4.47 times higher. At the same time, it effectively reduces the information loss caused by the conventional algorithms. The peak signal-to-noise ratio of the reconstructed image of the proposed algorithm is equivalent to or slightly lower than that of JPEG-LS and JPEG2000, but it has better visual quality and fidelity of the weak target area.