Camera Model Identification Based on Residual Extraction Module and SqueezeNet
Zeyu Zou, Yunxia Liu, Wenna Zhang, Yuehui Chen · 2019
Camera model identification is a hot topic in the field of image forensics. In this paper, a patch-level camera model identification method based on convolutional neural network is proposed. Firstly, inspired by pre-processing method in traditional camera model identification method, an automatic residual extraction module is designed in order to avoid subsequent convolutional neural network being affected by image content. Secondly, a modified SqueezeNet is proposed to extract the camera model related features within image patches. Finally, the effectiveness of the proposed method is verified under a strict patch-level evaluation protocol, which is designed based on the largest public image forensic Dresden database.