Vol 30, No 11 (2024)

The paper focuses on evaluating and improving the robustness of watermarking algorithms against DCT-filtering, a common image processing technique that can be used to remove embedded watermarks from cover images. The study examines the E_BLIND/D_LC watermarking algorithm, which operates in the spatial domain and is designed to embed a 1-bit message into an image. To evaluate robustness against DCT-filtering, the research utilizes software implementations of the watermarking algorithms, the dctdnoiz filter from FFmpeg, and the Flickr8k image database, which contains 8091 images. An extension to the E_BLIND/D_LC algorithm is proposed to enhance its robustness against DCT-filtering. The paper presents an experiment comparing the robustness of the extended algorithm with the original E_BLIND/D_LC. The results demonstrate that the extended algorithm improves robustness against DCT-filtering and also maintains better image quality, as measured by the PSNR metric, at equal levels of robustness in terms of the number of successful watermark detections after filtering. This suggests that the proposed extension effectively balances enhanced robustness against DCT-filtering with the preservation of image quality. The experiment confirms that the extended version outperforms the original algorithm in both detection reliability and image quality under DCT-filtering conditions. The paper includes links to the algorithm implementations on C++ programming language and the Python testing scripts, which can be useful for further research and practical applications in software development.

To ensure the quality of compilers, virtual machines and microprocessors, testing with the use of test programs is applied. Test templates, i.e. parameterized descriptions of test programs, are used to reduce the resources spent on test creation. For verification of test objects with different architectures, test templates testing the same functional properties can be used. A task of describing templates in an architecture-independent form and its customization for a target architecture arises. The paper proposes a method for creating and using architecture independent test templates. The method is as follows. Architecture dependent constructs that are needed to describe a test template are defined. Based on these constructs, an architecture independent test template is created. Then the template is customized to the architecture of the object under test by describing implementations of the architecture dependent constructs. Realizations of the constructs can be described by different sets of architecturally dependent operations, differing both by the type of data used and by the operations used. Different strategies for selecting the implementations to be used, such as full brute force, random selection, and others, can be applied in the test templates. Providing an interface as the architecture dependent constructs allows the use of third-party generators to create the architecturally independent templates. Rewriting the implementations of the constructs allows the test programs to be derived from the given templates for new architectures. This method was successfully applied to check optimizations in a virtual machine compiler.

The intensity of the use of remote work and learning tools has increased significantly during the pandemic and has led to the emergence of new tools and a significant increase in the capabilities of previously developed technology platforms. The introduction of modern technologies based on software control and complex virtualization for the creation of virtual laboratory complexes in educational institutions provides virtually unlimited opportunities for organizing individual and distance forms of practical classes. Modern learning management platforms that able to integrate with the database of university students makes it possible to evaluate the quality of students’ work with various types of course learning materials in real time. The article shows that the use of virtual communication infrastructures in connection to the Moodle learning management platform will provide end-to-end control of the theoretical and practical components of the educational process and will automate the processes of preparing and conducting practical classes, which will improve the quality of the educational processand the teacher efficiency.