Classification oriented embedded image coding | IEEE Conference Publication | IEEE Xplore

Classification oriented embedded image coding


Abstract:

This paper discusses the efficient compression of images to improve the classification associated with the decoded imagery. The set partitioning in hierarchical trees (SP...Show More

Abstract:

This paper discusses the efficient compression of images to improve the classification associated with the decoded imagery. The set partitioning in hierarchical trees (SPIHT) algorithm, an efficient wavelet-based progressive image-compression scheme, was originally designed to minimize the mean-squared error (MSE) between the original and decoded imagery. The image is first segmented at the encoder by an unsupervised method using a hidden Markov tree (HMT) mixture model in the wavelet domain. By using the kernel matching pursuits (KMP) method the recognition importance of each wavelet subband is estimated. By comparison using synthesized data, the compression and classification performance of the modified SPIHT algorithm is comparable to Bayes TSVQ, along with the advantages of fast speed and no requirement of codebook design and possibly transmission.
Date of Conference: 23-25 March 2004
Date Added to IEEE Xplore: 24 August 2004
Print ISBN:0-7695-2082-0
Print ISSN: 1068-0314
Conference Location: Snowbird, UT, USA

This paper discusses how to efficiently compress the image to improve the classification performance associated with the decoded imagery. The set partitioning in hierarchical trees (SPIHT) algorithm, an efficient wavelet-based progressive image-compression scheme, was originally designed to minimize the mean-squared error (MSE) between the original and decoded imagery. However, the MSE-based distortion measure is not in general well correlated with the image-recognition quality, especially at low bit rates. We modify the SPIHT algorithm to jointly reduce the MSE-based distortion and classification error.