Research of lossy image compression algorithms and the implementation of their combinations
Students Name: Tydniuk Dmytro Serhiiovych
Qualification Level: magister
Speciality: Computer Systems and Networks
Institute: Institute of Computer Technologies, Automation and Metrology
Mode of Study: full
Academic Year: 2022-2023 н.р.
Language of Defence: англійська
Abstract: Image compression – the use of special compression algorithms to reduce the size of image files by removing redundant elements from them. Compression reduces the time for their transmission over the network and between physical media. Also, the result will be a reduction in the space required for saving files, which is a very important indicator, because the large size plays a negative role when storing and transferring a large number of images, as it increases the cost of image storage. Data compression methods are aimed at reducing the volume of the original information stream using lossless or lossy transformations. Therefore, compression methods are divided into two categories: lossless compression and lossy compression. Digital image compression methods can be systematized according to their main characteristics: the accuracy of restoration, the symmetry of the main transformation and the type of transformation used. The object of research is lossy image compression algorithms. The subject of the research is the development of a proprietary lossy image compression algorithm based on combinations of such popular algorithms as JPEG and Haar’s DWT. The goal of work is to research popular image compression algorithms and develop your own. As a result of the master’s qualification work, a new image compression algorithm was created, which allows working with files of various formats, compressing both black and white and color images. Programmatically, this algorithm can be implemented on any of the modern operating systems. There is also the possibility of hardware implementation of this algorithm in the form of a FPGA core or a microcontroller. The scientific novelty is that a new lossy image compression algorithm is proposed, which is different from analogues present on the market and does not contain patent prohibitions, which will allow it to be used in many leading companies and industries that use image compression, while, that according to the numerical characteristics of the work results, the developed algorithm is at the level of its analogues, which are actively used in the modern world. Total volume: 101 pages, including 5 appendices totaling 18 pages, 22 figures, 15 tables. Keywords: lossy compression algorithms, lossless compression algorithms, JPEG, DCT, JPEG-2000, RLE, Huffman algorithm, Haar algorithm, DWT, C#