Hybrid feature-based analysis of video’s affective content using protagonist detection
作者：
Highlights：
• The protagonist with new definition is efficient to capture video emotional content.
• Temporal information is useful to improve the emotion recognition performance.
• A novel keyframe selecting strategy can extract typical protagonist-based frames.
• Combining optical flow and CNN can extract discriminative features in time domain.
• The proposed system can recognize the video emotion intelligently and automatically.
摘要
•The protagonist with new definition is efficient to capture video emotional content.•Temporal information is useful to improve the emotion recognition performance.•A novel keyframe selecting strategy can extract typical protagonist-based frames.•Combining optical flow and CNN can extract discriminative features in time domain.•The proposed system can recognize the video emotion intelligently and automatically.
论文关键词：Affective analysis,Video content analysis,Protagonist information,Convolutional neural network
论文评审过程：Received 12 February 2018, Revised 9 March 2019, Accepted 9 March 2019, Available online 11 March 2019, Version of Record 8 May 2019.
论文官网地址：https://doi.org/10.1016/j.eswa.2019.03.017
