Search

With the rapid growth of the multimedia technology, copyright protection has become a very important issue, especially for images. The advantages of easy photo distribution are discarded by their possible theft and unauthorized usage on different websites. Therefore, there is a need in securing information with technical methods, for example digital watermarks. This paper reviews digital watermark embedding methods for image copyright protection, advantages and disadvantages of digital watermark usage are produced. Different watermarking algorithms are analyzed. Based on analysis results most effective algorithm is chosen – differential energy watermarking. It is noticed that the method excels at providing image integrity. Digital watermark embedding system should prevent illegal access to the digital watermark and its container. Requirements for digital watermark are produced. Possible image attacks are reviewed. Modern modifications of embedding algorithms are studied. Robustness of the differential energy watermark is investigated. Robustness is a special value, which formulae is given further in the article. DEW method modification is proposed, it’s advantages over original algorithm are described. Digital watermark serves as an additional layer of defense which is in most cases unknown to the violator. Scope of studied image attacks includes compression, filtration, scaling. In conclusion, it’s possible to use DEW watermarking in copyright protection, violator can easily be detected if images with embedded information are exchanged.

Deep Fake is a technique for human image synthesis based on artificial intelligence. In this article is explored the problem of Deep Fake Video content and its detection. Has been gathered information about previous attempts, analyzed methods used by different researches and considered their actuality right now. Basing on results of the discovery was designed strategy to expose Deep Fake videos that combines previous detection methods with super-resolution algorithms. Results of the research were compared with expected, so recommendations and possible way of continuing developments were given.

This article discusses the problem of detecting dark spots in the urban video surveillance system. To identify shortcomings in the work of urban surveillance cameras, machine learning was used, which allowed to solve the problem of visualizing areas that are not in the camera's field of vision. Taking into account the fact that a wide range of video surveillance devices participated in the study: yard, porch video surveillance cameras and surveillance cameras in crowded places, a system was developed consisting of a database and an algorithm that allows you to request information only about the nearest video surveillance devices. This approach to the implementation of the main goal of the study contributed to the creation of an intuitive, user-friendly program that will help identify and eliminate dark spots in the urban video surveillance system.

Nowadays internet-trolls have big impact on other users, it interferes with comfortable use. Objective of the project is creating model for identifying provocative and fake comments. Science articles about detection of trolls by hand were searched for making the criteria of relevant comments selection. To achieve the goal there were created two artificial neural networks: definition of sarcasm and definition of sentiment analysis. The program result is datasets of troll comments and fake comments, statistic and diagrams of definition.