Feature Pyramid Deep Matching and Localization Network for Image Forensics | |
Kui Ye, Jing Dong, Wei Wang, Bo Peng, Tieniu Tan | |
2018 | |
会议名称 | APSIPA ASC 2018 |
会议日期 | 12-15 November 2018 |
会议地点 | Honolulu, Hawaii, USA |
出版者 | EI |
摘要 | Abstract--To advance the state of the art of image foren- sics technologies, a new formulation of splicing localization is proposed, which aims to obtain the masks for both the query and donor images for a pair of query(probe) image and potential donor image if a region of the donor image was spliced into the probe. The former Deep Matching and Validation Network(DMVN) addresses the problem with a novel end-to-end learning based solution. Inheriting the deep dense matching layer, we propose Feature Pyramid Deep Matching and Localization Network(FPLN), whose contributions are three folds. Firstly, in- stead of using just one feature map as in DMVN, FPLN utilizes a pyramid of feature maps with different resolutions w.r.t. the input image to achieve better localization performance, especially for small objects. Secondly, we add a fusion layer that fuses together all the features after deep dense matching layer, which not only takes full advantage of the correlation information between those features, but is also able to integrate two pathways in DMVN into just one simple pathway, simplifying the subsequent architecture. Lastly, we employ focal loss to address the imbalance problem, as the foreground area is usually much smaller than the background area. The experiments demonstrate the superior performance of our proposed method in detection accuracy and in localizing small tempered regions. |
DOI | 10.23919/APSIPA.2018.8659464 |
资助项目 | National Key Research and Development Program of China[2016YFB1001003] ; National Natural Science Foundation of China[U1636201] ; National Natural Science Foundation of China[U1636201] ; National Key Research and Development Program of China[2016YFB1001003] |
语种 | 英语 |
引用统计 | |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/25843 |
专题 | 智能感知与计算 |
通讯作者 | Kui Ye, Jing Dong, Wei Wang, Bo Peng, Tieniu Tan |
作者单位 | National Laboratory of Pattern Recognition, Center for Research on Intelligent Perception and Computing, Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China |
第一作者单位 | 中国科学院自动化研究所 |
通讯作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Kui Ye, Jing Dong, Wei Wang, Bo Peng, Tieniu Tan. Feature Pyramid Deep Matching and Localization Network for Image Forensics[C]:EI,2018. |
条目包含的文件 | 条目无相关文件。 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论