Deep Image Composition Meets Image Forgery
Eren Tahir, Mert Bal
TL;DR
This paper tackles the data bottleneck in image forgery detection by proposing an automated data-generation framework that uses image composition to create high-quality spliced images close to real manipulations. Leveraging the OPA dataset as a base, it employs MatteFormer for precise matting and Harmonizer for harmonization to produce 24,964 spliced images with a train/test split. Evaluation with a state-of-the-art detector shows reduced detectability on the generated data, indicating higher realism compared to existing datasets. The work provides a scalable pathway to richer, pixel-level forgery datasets that can generalize to broader manipulation types in future work.
Abstract
Image forgery is a topic that has been studied for many years. Before the breakthrough of deep learning, forged images were detected using handcrafted features that did not require training. These traditional methods failed to perform satisfactorily even on datasets much worse in quality than real-life image manipulations. Advances in deep learning have impacted image forgery detection as much as they have impacted other areas of computer vision and have improved the state of the art. Deep learning models require large amounts of labeled data for training. In the case of image forgery, labeled data at the pixel level is a very important factor for the models to learn. None of the existing datasets have sufficient size, realism and pixel-level labeling at the same time. This is due to the high cost of producing and labeling quality images. It can take hours for an image editing expert to manipulate just one image. To bridge this gap, we automate data generation using image composition techniques that are very related to image forgery. Unlike other automated data generation frameworks, we use state of the art image composition deep learning models to generate spliced images close to the quality of real-life manipulations. Finally, we test the generated dataset on the SOTA image manipulation detection model and show that its prediction performance is lower compared to existing datasets, i.e. we produce realistic images that are more difficult to detect. Dataset will be available at https://github.com/99eren99/DIS25k .