电视节目智能分类技术研究

he intelligent TV program classification is used for the classification of TV programs, including video data preprocessing, training and optimization of video classification model, testing the video classification model and analysis results. An intelligent TV program classification method based on NeXtNEW, DBoF and MoE is proposed. Video data preprocessing includes feature extraction, scene detection and scene segmentation. The training and optimization of video classification model include the training of baseline classification model, the training of feature adjustment model and the training of competing classification model. The test of video classification model includes the baseline classification model and the competing classification model. By comparing the recall rate and accuracy in the test results, it is concluded that when the number of clustering centers of DBoF model is 4096, the competing classification model is superior to the baseline classification model, and the classification effect is better.