Adaptive classification and subsampling for block transform image coding

Ih Hua Chang, Pao Chi Chang

Research output: Contribution to journalArticlepeer-review

Abstract

We present a block transform image coding system which preprocesses the image by adaptive subsampling based on the local activity levels. An input image is segmented into high/low activity blocks by combined L1 norm and mean difference classification methods. Low activity blocks are subsampled and encoded by conventional transform coding, such as JPEG, MPEG, and H.261, while high activity blocks are directly encoded by the transform coding. As a result, it yields good quality in both smooth and complex areas of the image at low bitrates.

Original languageEnglish
Pages (from-to)165-172
Number of pages8
JournalJournal of the Chinese Institute of Electrical Engineering, Transactions of the Chinese Institute of Engineers, Series E/Chung KuoTien Chi Kung Chieng Hsueh K'an
Volume4
Issue number2
StatePublished - May 1997

Fingerprint

Dive into the research topics of 'Adaptive classification and subsampling for block transform image coding'. Together they form a unique fingerprint.

Cite this