Analyzing and Detecting Digital Counterfeit Images using DenseNet, ResNet and CNN

This research study utilizes DenseNet architecture to detect copy-move forgeries in digital images. DenseNet’s architecture is utilized for efficient feature extraction, leveraging its dense connections to facilitate improved feature learning. Initial preprocessing tasks are performed to prepare the images, followed by patch extraction and feature extraction using DenseNet. Additionally, a ResNet model is trained on a dataset containing authentic and manipulated images, which is then utilized to identify forgeries in unseen images. Further enhancements in results can be achieved through post-processing techniques. The primary goal of this research is to enhance the precision and reliability of copy-move forgery detection in digital images. This objective is pursued by developing and implementing a DenseNet-based methodology for feature extraction and learning. By leveraging the inherent strengths of DenseNet architecture, the system aims to achieve higher performance levels in detecting copy-move forgeries compared to traditional methodologies. The ultimate aim is to drive progress in the field of image forensics and strengthen the integrity of digital visual content across various domains.