Multistage Vector Quantization of Image by Using Clustering Interpolation in DCT Domain
Abstract
An adaptive image coding scheme, called multistage vector quantization (MSVQ) is proposed. A new algorithm of MSVQ uses clustering interpolation in DCT domain and enables vector quantization of large image blocks with tolerable encoding complexity. Their mean luminance values are efficiently removed by the spline interpolative vector quantizer using a small number of bits. It can achieve a very high compression rate for low detailed images without complicated classification of transform blocks and subvector construction.
Keywords
vector quantization, clustering interpolationPersistent identifier
http://hdl.handle.net/11012/58373Document type
Peer reviewedDocument version
Final PDFSource
Radioengineering. 1997, vol. 6, č. 4, s. 10-13. ISSN 1210-2512http://www.radioeng.cz/fulltexts/1997/97_04_03.pdf
Collections
- 1997/4 [5]