Color image splicing localization algorithm by quaternion fully convolutional networks and superpixel-enhanced pairwise conditional random field

Color image splicing localization algorithm by quaternion fully convolutional networks and superpixel-enhanced pairwise conditional random field

Recently, fully convolutional network (FCN) has been successfully used to locate spliced regions in synthesized images. However, all the existing FCN-based algorithms use real-valued FCN to process each channel separately. As a consequence, they fail to capture the inherent correlation between color...

Full description

Bibliographic Details
Main Authors: Beijing Chen, Ye Gao, Lingzheng Xu, Xiaopeng Hong, Yuhui Zheng, Yun-Qing Shi
Format: Article
Language:English
Published: AIMS Press 2019-07-01
Series:Mathematical Biosciences and Engineering
Subjects:
splicing localization
splicing detection
quaternion
fully convolutional network
conditional random field
Online Access:https://www.aimspress.com/article/10.3934/mbe.2019346?viewType=HTML

Internet

https://www.aimspress.com/article/10.3934/mbe.2019346?viewType=HTML

Similar Items