一种无参考图像的块效应检测算法
No-reference Blockiness Detection Algorithm
基于块的视频编码方法,在低比特率时解码恢复出的图像的块边界上会出现比较明显的块效应,降低了图像的视觉质量,因此需要对存在块效应的图像进行去块效应滤波。本文提出了一种新的无参考图像的块效应检测算法,该算法使用三个限制条件来区分块效应与图像的真实边缘和纹理,以更准确的检测块效应的存在性、块边界的位置和块效应的程度。基于一帧图像的块效应评估值可以用于调节去块滤波的强度,实现帧级的自适应。仿真实验结果证明该方法的有效性。本算法同时具有计算复杂度低、适用于实时处理等优点。
Video coding methods based on the block will bring about blocking artifacts in reconstructed picture when applied in low bit rate environment. The artifacts may lead to the decline of visual image quality, so it is necessary for the blocky images to de-block. The paper proposes a new no-reference blockiness detection algorithm. The algorithm uses three constraints to distinguish block boundaries from real edges and textures, so as to detect more accurately the existence of blockiness, the positions of block boundaries and the extent of blockiness. The blockiness evaluation result can be used to tune the strength of the de-blocking filter which is adaptive at the frame level. The experimental results prove the effectiveness of the algorithm. The technique can meet the need of real time process for its simpleness.
郑旭
电子技术应用
图像压缩块效应检测无参考图像去块效应
Image CompressionBlockiness DetectionNo-referencee-blocking
郑旭.一种无参考图像的块效应检测算法[EB/OL].(2009-02-19)[2025-08-02].http://www.paper.edu.cn/releasepaper/content/200902-988.点此复制
评论