The new algorithm achieves very close quality compared to full search but with several hundred times speedup.
新算法与全搜索块匹配算法相比,在重建图像质量方面是非常接近的,但却获得了几百倍的提速。
EPZS achieves good search performance, maintaining the SNR almost the same of full search and increasing search speed about 200 times.
EPZS实现了很好的搜索性能,保持了与全搜索差不多的信噪比,搜索速度提高了200倍左右。
Experimental data on full search block matching algorithm indicates that the proposed method keeps the accuracy while the computation load is reduced to 8.72%.
以全搜索块匹配算法为例的实验表明,该算法在保持精度不变的前提下,运算量最终减少到8.72%。
The test result shows that the reduction of computation is about 90% in the encoding time against full search method, and a little precomputation and extra storage are required.
测试结果表明,相对于穷尽搜索方法,计算量得到明显的降低,计算时间减少约90 % ,同时只需要很少的预先计算量和额外存储量。
Experimental results show that our algorithm can reduce the computational-complexity to 3% of fast full search motion estimation algorithm, with negligible PSNR loss and bit rate increase.
试验结果证明,该算法能够将计算复杂度降低到快速全搜索运动估计算法的3%,而PSNR和码率几乎与快速全搜索算法得到的结果相当。
In this paper, a quasi full search method for the improved motion estimation algorithm is proposed.
提出了一种称为准全搜索法的块匹配改进算法。
Full search algorithm searches each point covering the full search window, and too much computation leads to very slow search speed which places great impact on its application.
全搜索算法搜索遍及搜索窗每一个点,计算量太大,搜索速度严重影响了其应用。
Experiments show that the algorithm can obtain almost the same PSNR as full search (FS) at a low computation cost, and it excels three-step search (3ss) algorithm in searching velocity and effect.
实验结果表明,该算法可以较小的搜索代价取得与全搜索相当的效果,在搜索速度与搜索效果方面优于三步法(3ss)。
Simulation result shows that the proposed algorithm has a prediction quality near to the Full Search, and reduces the search points in about 10% relative to the Hexagon Search Algorithm.
仿真结果表明,在重建图像质量接近全搜索的同时,新算法比六边形算法减少了约10%的搜索点数。
The larger the codebook size or the vector dimension is, the higher the computational complexity of the full-search VQ is.
码书大小和矢量维数越大,穷尽搜索矢量量化编码的计算复杂度就越高。
The larger the codebook size or the vector dimension is, the higher the computational complexity of the full-search VQ is.
码书大小和矢量维数越大,穷尽搜索矢量量化编码的计算复杂度就越高。
应用推荐