Composite algorithms for screen video key frames compression

Video compression is one of the most time-taking problems, which are solved not only by professionals, but also by ordinary users. Screen video is one of the video data types, which must be compressed in real-time mode. Often it is necessary to compress screen video in background mode at that. There is a necessity in fast image compression algorithms, because during video compression key frames are encoded independently. Screen video frames are referred to discrete-tone images. As a rule algorithms without information loss are used to compress such images, because while compressing such images even a small information loss percent can result in significant visual image degradation. RLE and family of LZ algorithms can be referred to such algorithms. LZO is actively used in screen video compression (for example, it is used in freeware screen video recorder CamStudio). Composite algorithm for screen video key frames compression is introduced in this paper. This algorithm uses hybrid algorithm, developed by the author earlier, on the first step and zlib library on the second step. The paper also contains results of practical comparison with composite algorithm, which uses hybrid algorithm on the first step and LZO algorithm on the second step, and with several other algorithms. Three types of screenshots were used for testing: 1. Pictures, typical for Windows XP (1Q pieces); 2. Pictures with text (8 pieces); 3. Pictures with graphics, diagrams (1Q pieces). Introduced in this paper composite algorithm, based on hybrid algorithm and zlib, confirmed its effectiveness. This composite algorithm demonstrates higher degree of compression (at the average 23,5 %o more for pictures, typical for Windows XP comparing to zlib level 6). Meanwhile, introduced composite algorithm demonstrates higher compression speed (at the average 4 times faster comparing to zlib level 9 for pictures with text). Therefore, introduced combined algorithm can be used in practice for screen video frames compression.

Keywords

Authors

References