Arbitrary-sized Image Training and Residual Kernel Learning: Towards Image Fraud Identification
This work addresses image fraud detection, a critical issue for security and forensics, by introducing a novel training method to handle arbitrary-sized images, though it appears incremental in its approach to noise residual preservation.
The paper tackled the problem of image fraud identification by preserving original noise residuals, which are damaged by resizing in deep learning, and achieved state-of-the-art results, particularly for images with small tampered regions or unseen tampering distributions.
Preserving original noise residuals in images are critical to image fraud identification. Since the resizing operation during deep learning will damage the microstructures of image noise residuals, we propose a framework for directly training images of original input scales without resizing. Our arbitrary-sized image training method mainly depends on the pseudo-batch gradient descent (PBGD), which bridges the gap between the input batch and the update batch to assure that model updates can normally run for arbitrary-sized images. In addition, a 3-phase alternate training strategy is designed to learn optimal residual kernels for image fraud identification. With the learnt residual kernels and PBGD, the proposed framework achieved the state-of-the-art results in image fraud identification, especially for images with small tampered regions or unseen images with different tampering distributions.