The comparison and analysis of extracting video key frame

International Conference on Computer Information and Automation Engineering

计算机科学;运输工程

Ouyang, S.Z.^1 ; Zhong, L.^1 ; Luo, R.Q.^2

Wuhan University of Technology, Research Interests Include Data Mining and Image Processing, China^1

Wuhan University of Technology, Research Interests Include Picture Processing and Computer Vision, China^2

关键词: Comparison and analysis;    Frame differences;    Information entropy;    Key-frame extraction;    Key-frames;    Mutual informations;    Mutual trust;    threshold;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/359/1/012010/pdf
DOI  :  10.1088/1757-899X/359/1/012010

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

Video key frame extraction is an important part of the large data processing. Based on the previous work in key frame extraction, we summarized four important key frame extraction algorithms, and these methods are largely developed by comparing the differences between each of two frames. If the difference exceeds a threshold value, take the corresponding frame as two different keyframes. After the research, the key frame extraction based on the amount of mutual trust is proposed, the introduction of information entropy, by selecting the appropriate threshold values into the initial class, and finally take a similar mean mutual information as a candidate key frame. On this paper, several algorithms is used to extract the key frame of tunnel traffic videos. Then, with the analysis to the experimental results and comparisons between the pros and cons of these algorithms, the basis of practical applications is well provided.