Video forgery is gaining popularity in digital networks making it imperative to develop viable methods to combat altering. This research work describes the progression of a new model which employs a Resnet50 model in detecting forged videos. The model integrates multi-scale feature extraction which helps it to detect fine content alterations. The proposed method achieves an accuracy of 89.7% in locating and marking the occurrence of tampered regions on videos. Therefore, such processes underline the importance of the efficient detection methods in digital forensics because they provide a great deal of help in addressing the issues of video frames alteration without raising the digital credibility concerns.