Hybrid image splicing detection: Integrating CLAHE, improved CNN, and SVM for digital image forensics

Abstract

This article introduces a novel hybrid method for the detection of image splicing forgery (ISF) that integrates an improved convolutional neural network (CNN), support vector machine (SVM) classifier, and contrast-limited adaptive histogram equalization (CLAHE). The imperceptibility of counterfeit images has made detection a challenge, as the increasing accessibility of image editing applications has resulted in a surge in amateur image manipulation. The proposed methodology employs CLAHE to enhance the extraction of hidden features that forgery has obscured. The improved CNN employs sophisticated feature extraction techniques to achieve superior classification accuracy without the necessity of custom algorithms. Furthermore, SVM is incorporated due to its exceptional processing speed and efficiency. The objective of this hybrid framework is to address the constraints of current deep learning models in terms of computational efficiency and accuracy, thereby demonstrating substantial enhancements in performance metrics for image splicing forgery detection (ISFD). The findings suggest that the proposed system effectively differentiates between authentic and manipulated images, offering an effective solution to the challenges of image splicing forgery.