Enhancing Medical Imaging Trust with an Efficient Deep Learning System for Forgery Detection

Abstract

The integrity of medical images is critical to accurate diagnosis and treatment, but the proliferation of sophisticated editing tools has heightened the risk of forgery, compromising the trustworthiness of medical imaging. This paper introduces an efficient deep learning system designed to safeguard medical imaging by identifying alterations such as splicing, copy-move, and retouching forgeries. Our proposed solution utilizes a convolutional neural network (CNN) architecture enhanced with attention mechanisms, optimized to detect subtle changes in medical images. The development process involved meticulous data preparation, augmentation, and the implementation of advanced mathematical models to refine the system's ability to differentiate between authentic and manipulated images. Evaluation was conducted on a diverse dataset encompassing various imaging modalities like MRI, CT scans, and X-rays, with both genuine and tampered instances. The proposed system delivered superior performance, achieving an accuracy of 97.5%, a precision of 96.8%, a recall of 98.2%, and an F1 score of 97.5<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, significantly outperforming existing methods. These results highlight the system's robustness in distinguishing real images from forged ones, thanks to its attention mechanism that allows the model to focus on key areas within the images. Ultimately, this system offers a highly accurate and efficient tool for enhancing the security of medical imaging, reinforcing the potential of deep learning in tackling healthcare challenges and ensuring reliable diagnostics and treatment planning.

Ähnliche Arbeiten