A vision-based fully automated approach to robust image cropping detection
作者：
Highlights：
• This work presents an automated approach for detecting asymmetrically cropped images.
• The shift of the image center w.r.t. the principal point is used to assess cropping.
• A novel metric based on a Monte Carlo analysis is proposed to evaluate cropping.
• Experiments show the effectiveness and robustness of the proposed method and metric.
• Heuristic criteria are also introduced to discard intractable images at run time.
摘要
•This work presents an automated approach for detecting asymmetrically cropped images.•The shift of the image center w.r.t. the principal point is used to assess cropping.•A novel metric based on a Monte Carlo analysis is proposed to evaluate cropping.•Experiments show the effectiveness and robustness of the proposed method and metric.•Heuristic criteria are also introduced to discard intractable images at run time.
论文关键词：Multimedia forensics,Robust computer vision,Cropping detection,Image content analysis
论文评审过程：Received 16 March 2019, Revised 16 July 2019, Accepted 3 September 2019, Available online 23 September 2019, Version of Record 25 September 2019.
论文官网地址：https://doi.org/10.1016/j.image.2019.115629
