This document summarizes a presentation on detecting digital image forgery using salient keypoints. It introduces common types of image forgery and clues that reveal forgery. A framework is proposed that selects salient keypoints using distinctiveness, detectability, and repeatability to reduce keypoints and detect copy-move forgery. The approach uses SIFT and KAZE features and achieves promising results on standard datasets, outperforming other methods with lower false positive rates and higher precision and F1 scores. Future work could detect other forgery types and develop more robust detection algorithms.