Software for editing images has advanced alarmingly as it has become available to everyone. This causes a lot of problems with forgery. One of the most common types of cheating is the splicing attack. Siamese Networks are useful tools for pattern recognition, especially in the case of applications involving comparison or matching of different data like pictures or texts. The paper provides a deep learning approach based on Siamese Network architecture to detect image splicing better. It is when a piece of one image is cut out and pasted into another. It is a serious issue in image forgery. Siamese neural network (SNN) identifies distinctive features of certain parts of the image and uses the features to pinpoint tampered areas. The design has two similar subnetworks with standard weights and image features to inspect the picture blocks and recognize the altered regions. Following that, photo segmentation and detection of the changes take place using a segment anything model (SAM). The effectiveness of our approach is verified on datasets containing splicing attacks. The result has an accuracy of 98.5% and a precision of 97.7%. This innovative work employs SAM and Siamese Networks to segment the image accurately and quickly and precisely detect image tampering by the Siamese network.