Image Compression for Storage and Transmission of Digital Images
Vladimir Cherkassky, Xuhao He, Jie Shao
Report no. Mn/DOT 2001-14
Topics: Traffic Modeling and Data
This project researched image compression methods for storage and transmission of digital images at the Minnesota Department of Transportation (Mn/DOT). Researchers compared the performance of several commercial and research methods for image compression based on the "typical" image provided by the Mn/DOT Office of Land Management. They also surveyed some new image compression methods based on wavelet thresholding. The report details the analyses and comparisons and includes recommendations. Researchers choose MrSID, a commercial software package for image compression, as a suitable method for the needs of Mn/DOT's Office of Land Management. MrSID uses a wavelet transform-based algorithm to achieve both the efficient storage and retrieval of large digital images. Its main practical advantages include improved utilization of storage and transmission resources and a multi-resolution browsing capability. MrSID can selectively decompress a portion of an image by zooming at different levels of detail. Researchers implemented a demo website, which can be browsed easily and implements the zooming function. The web site address is http://rocky.dot.state.mn.us/research.