基于编码量化参数调节的图像清晰化处理  

Image Clarity Processing Based on Encoding Quantization Parameter Adjustment

作　　者：孙令翠 冯辉宗[2] SUN Ling-cui;FENG Hui-zong(College of Computer and Internet of Things,Chongqing Institute of Engineering,Chongqing 400056,China;College of Automation,Chongqing University of Posts and Telecommunications,Chongqing 400065China)

机构地区：[1]重庆工程学院计算机与物联网学院,重庆400056 [2]重庆邮电大学自动化学院,重庆400065

出　　处：《计算机仿真》2022年第4期180-184,共5页Computer Simulation

基　　金：基于光学字符识别的高精度阅卷方法研究(KJQN202001902)。

摘　　要：从人类视觉角度出发研究融合视觉感知特性的图像失真校正方法,提升视频图像中的图像清晰程度,改善图像失真问题。定义视觉感知因子,综合考虑视觉感知过程中图像纹理、亮度以及空域活动性的感知特性,将当前编码树单元作为单位融合视频内容自适应调整拉格朗日乘子,按照量化参数和拉格朗日乘子之间存在的关系,使用编码量化器实现量化参数动态调节,进一步去除视频图像中的视觉感知冗余,实现图像失真校正。经过实际视频序列图像分析,在调节参数为0.6的情况下,编码与失真平衡效果最好,上述方法中由于使用双边滤波器保留图像高频纹理细节,确保图像失真校正具有更好的效果,同时,上述方法还校正图像细节失真问题,使得图像更加清晰。Based on human vision, this paper studied the image distortion correction method integrating visual perception characteristics in order to improve the image clarity and image distortion in video images. Visual perception factors were defined. In the process of visual perception, the perceptual characteristics of image texture, brightness and spatial activity were combined. The existing coding tree unit was used as a unit to fuse video content to adaptively adjust Lagrange multipliers. According to the relationship between quantization parameters and Lagrange multipliers,the coding quantizer is used to realize the dynamic adjustment of quantization parameters, further remove the visual perception redundancy in the video image and realize the image distortion correction. The experimental results show that this method has excellent image distortion correction effect owing to the use of bilateral filter. The balance effect between coding and distortion is the optimal(the adjustment parameter is 0. 6).

关 键 词：视觉感知特性 图像 失真校正 拉格朗日乘子 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]