A Novel Method for Color Forged Image Detection
Ze Yang, Rui Guo, Xuejuan Chen, Zhihua Xiang, Yuying Liang · 2021 IEEE 5th Information Technology,Networking,Electronic and Automation Control Conference (ITNEC) · 2021
Digital images permeate almost all areas of our lives, mainly in news media, scientific discoveries, medical imaging, and judicial evidence. However, in recent years, due to the wide application of deep learning in image processing technology, these technologies or software have not only brought convenience to people, but also made it easier for people to forge or tamper with digital images without leaving any traces. The authenticity of digital images has been affected. A serious threat to modify and threaten. These forged or tampered images will bring serious threats to judicial justice, social stability, and the medical industry, and cause huge negative effects. Therefore, this article proposes an authenticity detection algorithm for generating color forged images based on deep learning. The corresponding color channel features of the real and forged image datasets are extracted and FIsher encoded, respectively, and the encoded color channel features are used to train the SVM model. Experiments prove that our proposed method achieves better results in detecting image color tampering.