Thesis - Open Access
Master of Science (MS)
Robert G. Finch
Data compression is essential for transmission and storage of images due to the amount of image data acquired every day. Many techniques have been suggested in the literature in the past two decades for the purpose of image compression. Vector quantization is a widely used scheme for image compression, and it is theoretically the best method since, according to Shannon’s source coding theorems, better results can be achieved by quantizing vectors instead of scalars . A vector quantizer is a system that maps a vector space into a binary sequence. This sequence is transmitted or stored using fewer bits, achieving thus the compression desired. Many variations of vector quantization exist. The pairwise nearest neighbor clustering algorithm was implemented and tested at the EROS Data Center in Sioux Falls, SD. Compression ratios of 8 to 32 were achieved, with an encoding bit rate of 0.15 to 1.15. Reconstructed image quality was studied and encoding and decoding run times were recorded.
Library of Congress Subject Headings
Data compression (Telecommunications)
Data compression (Computer science)
Number of Pages
South Dakota State University
Karam, Walid Philip, "Application of Image Data Compression Using Vector Quantization" (1990). Electronic Theses and Dissertations. 5336.