Improved model for image tampering monitoring based on fast-RCNN
Jiaqi Chen, Jiasheng Chen, Weitong Wang, Yunlong Zhu · 2023
Image tampering detection is a method to detect if an image has been edited or moved. The most common tampering methods are copy, paste and splicing. The current techniques are divided into two main categories: traditional feature extraction-based tampering detection technology and convolutional neural network-based tampering detection technology. In this paper, we train images based on a dual-stream faster R-CNN network to detect the operation images given end-to-end in the tempered area. We replace VGG16 in the convolutional layer model with VGG19, smaller convolutional kernels, smaller pooling kernels and deeper layers are used to make wider feature mapping, learn the appropriate features from a large group of samples more efficiently. And in that case, we avoid unnecessary complex feature extraction process. We also demonstrate our approach through a set of experiments, and the experimental results have shown the effectiveness of using VGG19 instead of VGG 16.