Canonical image selection based on human affects in photographic images
作者：
Highlights：
• Affect based canonical image selection is suggested for summarizing a scene topic.
• PAM based PLSA learning is proposed to transfer visual space into affective space.
• Cluster-ranking model is presented to find a diverse set of representative images.
• Thereafter, the high-ranked images are selected from the top-ranked clusters.
• The experiments showed that the proposed method outperforms other baselines.
摘要
•Affect based canonical image selection is suggested for summarizing a scene topic.•PAM based PLSA learning is proposed to transfer visual space into affective space.•Cluster-ranking model is presented to find a diverse set of representative images.•Thereafter, the high-ranked images are selected from the top-ranked clusters.•The experiments showed that the proposed method outperforms other baselines.
论文关键词：Canonical image selection,Human affects,Probabilistic affective model,Ranking model,PLSA learning
论文评审过程：Received 14 March 2016, Revised 12 July 2016, Accepted 26 August 2016, Available online 6 September 2016, Version of Record 2 October 2016.
论文官网地址：https://doi.org/10.1016/j.imavis.2016.08.013
