05557738

a

05557738 Tsai, Cheng-Fa Lin, Yu-Jiun LISA: Image compression scheme based on an asymmetric hierarchical self-organizing map. Yu, Wen (ed.) et al., Advances in neural networks -- ISNN 2009. 6th international symposium on neural networks, ISNN 2009, Wuhan, China, May 26--29, 2009. Proceedings, Part III. Berlin: Springer (ISBN 978-3-642-01512-0/pbk). Lecture Notes in Computer Science 5553, 476-485 (2009). 2009 Berlin: Springer EN image compression vector quantization SOM HSOM

doi:10.1007/978-3-642-01513-7_52

Summary: A Kohonen network, also called Self-Organizing Map (SOM), is a competitive learning network, and is appropriate for solving an image compression problem owing to its ability to generate high-quality compressed images. However, SOM has a large computation cost, making it impractical due to a lengthy training process. Hence, the Hierarchical Self-Organizing Map (HSOM) had been presented and found to reduce computation cost. Although a hierarchical architecture speeds up SOM, HSOM is still not practical enough because of a high compression cost. Therefore, this investigation employs a hybrid scheme to increase the efficiency and effectiveness of HSOM. Simulation results reveal that the proposed algorithm is much more efficient and effective than other algorithms, such as LBG, SOM, and HSOM.