一种新的低功耗图像及视频编解码算法

提出一种新的低功耗图像及视频编解码算法。该算法主要基于矢量量化，认为编码算法的质量和功耗地码本尺寸的大小，通过采用小尺寸码本，降低算法所需要的内存数量，从而降低功耗。编码时，利用分形理论中的同构变换计算虚拟码本，弥补由于采用小码本造成的图像质量损失，并使编码过程较少依赖于码本内存。编解码结果与全搜索型矢量量化算法相比，在不损失图像质量的前提下，可以极大地降低编解码功耗。

A novel algorithm for low power coding and decoding of image and video

A novel VQ based algorithm for low power image and video coding is proposed in this paper. This algorithm considers that power consumption and image quality in coding process depend on the codebook size, uses small codeboods to reduce its memory requirements, and adopts isometry transformations from the fractal theory to compute virtual codebook in order to improve image quality. A number of additional codewords are generated through computations, and the coding procedure becomes more computation based than memory based. In comparison to classical full search vector quantization, the algorithm leads to power savings without image quality reduction.