Key frame screening method based on interested target distribution

Abstract

The invention discloses a key frame screening method based on interested target distribution, and belongs to the technical field of image processing. The method comprises the following steps: carrying out feature extraction on each video frame image by adopting a plurality of feature extraction modes, and carrying out normalization processing on extracted feature vectors; calculating the feature distance between adjacent frames under each feature vector, and obtaining the difference between the adjacent frames through the weighted sum of all the feature distances; based on the difference curve between adjacent frames, realizing shot segmentation according to local adaptive dual thresholds, and performing target detection processing on each video frame image based on a neural network; detecting the obtained video frame in the shot, and obtaining the type and position of a target and the area of a detection frame; further segmenting the video frame image in the shot into sub-shots according to the target distribution and quantity difference in the video frame based on the target detection result; and selecting a frame with the highest target information richness in the sub-shot boundary as a key frame. The invention can be applied to monitoring videos and movie and television videos.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a key frame screening method based on the distribution of interested objects. Background technique [0002] With the development of my country's urban intelligence and Internet of Things technology, surveillance video systems have been used in various industries. Although surveillance video has the advantages of strong expressiveness, large amount of information, and vivid images, its shortcomings such as large changes, huge data volume, and low level of abstraction make its management and retrieval quite difficult. With the rapid growth of video data (about every 10 years, the total amount of video data will increase by 50 times), the demand for human resources and memory must be higher and higher. With the advancement of technology, the emergence of technologies such as key frame extraction will provide a scientific solution to this problem. [0003] The ...

AI Technical Summary

Problems solved by technology

However, these algorithms also have shortcomings such as not being able to adaptively set the key frame extraction threshold, low extraction flexibility, and too one-sided use of video image information. Secondly, traditional key frame extraction is based on the underlying features of the image, which cannot reflect video very well. The target information, the extracted key frame cannot contain the target object (interesting content) well, and it is easy to cause problems such as missing targets.

Images