首页> 外文会议>Chinese conference on pattern recognition and computer vision >Retinex Based Flicker-Free Low-Light Video Enhancement
【24h】

Retinex Based Flicker-Free Low-Light Video Enhancement

机译:基于Retinex的无闪烁微光视频增强

获取原文

摘要

Videos captured in low light environment tend to be poor visual effect. To get better visual experience, a video enhancement algorithm based on improved center-surrounded Retinex and optical flow is proposed in this paper, which contains intra-frame brightness enhancement and inter-frame brightness continuity. In intra-frame brightness enhancement, reflection of each frame is estimated by adjusting the illumination using a weight factor, so that bright illumination is compressed to obtain a reflection with approximately uniform illumination. Then logarithmic image processing subtraction (LIPS) is adopted to enhance its contrast. To maintain inter-frame brightness continuity, the background and brightness changes of adjacent frames are measured using optical flow and just noticeable difference (JND) threshold, respectively. If the background and average brightness change little, their reflection brightness is almost the same, so LIPS parameter of previous frame is applied to current frame. Otherwise, current frame will be updated by calculating its own parameter. Experimental results demonstrate that proposed algorithm performs well in brightness continuity and detail enhancement.
机译:在弱光环境下拍摄的视频往往会产生较差的视觉效果。为了获得更好的视觉体验,本文提出了一种基于改进的中心环绕的Retinex和光流的视频增强算法,该算法包含帧内亮度增强和帧间亮度连续性。在帧内亮度增强中,通过使用权重因子调整照明来估计每个帧的反射,使得明亮的照明被压缩以获得具有近似均匀照明的反射。然后采用对数图像处理减法(LIPS)来增强其对比度。为了保持帧间亮度连续性,分别使用光流和仅明显差异(JND)阈值来测量相邻帧的背景和亮度变化。如果背景和平均亮度变化不大,则它们的反射亮度几乎相同,因此将前一帧的LIPS参数应用于当前帧。否则,将通过计算自己的参数来更新当前帧。实验结果表明,该算法在亮度连续性和细节增强方面表现良好。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号