Piracy Estimation and Detection in Movies and OTT Series using Similarity Functions and Deep Learning Techniques

The growth of digital media has made it increasingly tough to prevent piracy; hence, it has been a major challenge to the film industries and OTT (over the top) platforms. This system uses advanced technological solutions to put up a fight against piracy. This system used a convolutional neural network (CNS), especially the VGG16 (Visual Geometry Group 16) model, to acquire features from video frames efficiently. These features from deep learning allow the system to identify and match content within large datasets. The system calculates inter-frame similarities through the computation of cosine similarities. This is an interlocking method that will make its piracy detection more accurate and reliable. The system will then identify and compare pirated content with a much higher level of confidence by using a cosine similarity function with the strong feature extraction capability of the VGG16 model. Extensive datasets are used that contain highly diversified movie types and OTT shows for broad testing and verification purposes. Results did prove the efficiency of the effort in terms of pirated content, which otherwise must have required a scalable and effective solution within the entertainment industry. The effort further helps generalize formulation measures for the protection of intellectual property while also trying to bring down piracy.