A Review On Deep Learning Techniques For Image Forgery Detection
Ashutosh Kumar Pandey, Pritee Parwekar, Ashish Kumar Chakraverti · 2022 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS) · 2022
The images are very popular medium now days for communication either in social media or by e-communication. It is very important that image circulated for communication must be original and authentic, so detection of forged images is significance. Images are forged mainly due to the various technological, moral, and judicial connotations associated with advance image editing software, it becomes very difficult to differentiate between original image and tampered image. Traditional methods for image forgery detection mostly use handcrafted features. The problem with the traditional approaches of detection of image tampering is that most of the methods can identify a specific type of tampering by identifying a certain feature in image. Nowadays, deep learning techniques are being used for image tampering detection. These methods reported better performance than traditional methods because of their capability of extracting complex features from image. In this paper, we present a detailed survey of deep learning-based techniques for image forgery detection, outcomes of survey in form of analysis and findings, and details of publicly available image forgery datasets.