Exploring Pseudo-Zernike Moments for Effective Image Forgery Detection

This research explores the synergistic integration of Pseudo-Zernike moments and machine learning techniques to enhance the accuracy and robustness of image forgery detection. Leveraging the discriminative power of Pseudo-Zernike moments in capturing shape and texture features, the study evaluates various machine learning models including K-Nearest Neighbors (KNN), Decision Tree (DT), Support Vector Machine (SVM), Random Forest (RF), Naive Bayes (NB), and an Ensemble Extra Trees model. Through comprehensive experimentation and evaluation on a dataset containing both authentic and manipulated images, the Ensemble Extra Trees model emerges as the most accurate and reliable approach for detecting image forgeries. The ensemble nature of Extra Trees, coupled with its ability to mitigate overfitting and reduce variance, enables it to effectively discern subtle alterations and manipulations within images, showcasing its potential for robust image forgery detection in digital image forensics.