Improved transform based image compression methods

Abstract

A Modified Discrete Wavelet Transform (DWT) and Hybrid image compression methods which use integral transformations are introduced in this paper. Discrete Cosine Transform (DCT) as implemented in the standard JPEG compression and DWT for its useful multiresolution properties are used first. A Modified DWT method with variable thresholds used to discard some coefficients is suggested. A mathematical model consisting of the proposed Hybrid Method that takes advantage of both the DCT and Modified DWT is then introduced. A new approach to truncate transform coefficients is followed in the proposed method to obtain wide range of bit rates in bits per pixel (bpp). The Modified DWT and proposed Hybrid compression methods are compared to the JPEG and JPEG2000 compression standards. The performances of the suggested methods show clear improvement (about 7dB) in Peak Signal-to-Noise Ratio and better Mean Structural Similarity (MSSIM) of the decompressed image (about 22%) compared to the standard JPEG2000 method when operating at the same bit rate.

A Comparative Study on Image Compression Method Using Stand-Alone DWT Techniques and Hybrid of DWT with Huffman Coding Technique for WSN Application

“…Discrete Wavelet Transform (DWT) is categorized in lossy compression method which has the ability in reducing the matrix size of image [9] with low amount of resolution loss [10]. Commonly, a lossy compression usually have some loss during compression, however in DWT technique, it offers less computational complexity rather than sacrificing the quality of image [11].…”