Tampered Image Detection Using Texture and Co...

Tampered Image Detection Using Texture and Color Based Feature Extraction: A Comparative Analysis with Machine Learning Classifiers

Author Name : Aziz Makandar, Shilpa Kaman

ABSTRACT This study presents a comprehensive approach to image forgery detection using texture and color feature extraction techniques. Specifically, Local Binary Patterns (LBP) and Gray Level Co-occurrence Matrix (GLCM) are utilized for texture analysis, while color histograms are computed in HSV and YCbCr color spaces to capture color information. The extracted features are then classified using Random Forest (RF) and Support Vector Machine (SVM) algorithms to detect tampered images. The performance of these classifiers is evaluated on two widely used datasets, MICC-F220 and CASIA 1.0, using accuracy, precision, recall, and F1 score as metrics. Experimental results indicate that Random Forest consistently outperforms SVM across both datasets, with the highest accuracy achieved using color histograms in the YCbCr space combined with LBP features. This study underscores the effectiveness of combining texture and color features for image forgery detection and highlights the importance of choosing appropriate feature extraction techniques and classifiers in digital image forensics.