An innovative web-based system for high lossless compression and fast, interactive transmission of visible human color images. The Third Visible Human

This paper describes a system uniquely designed for the management and distribution of the Visible Human (VH) cryosectioned color images of the male and the female. Moreover, this design approach is applicable to the management and distribution of other large image datasets when optimized to their specific characteristics. This research work includes the design and implementation of a robust lossless compression algorithm to store and transmit VH images over the Internet, while also allowing easy access to the images through the interactive selection of any section of the database for downloading and viewing in the users’ machines, accomplished by an interactive Java applet. Our system design is based on a color transformation from RGB to HSI to facilitate an automated extraction of an anatomic object from the background yielding a segmented image prior to compression. This unique two-step segmentation process, followed by lossles coding using an Adaptive Arithmetic Coding (AAC) model, yields an average data reduction of 7.8:1 for both male and female datasets and as high as 16:1 for specific slices. Reconstruction of the compressed images is achieved by applying the inverse AAC algorithm. The final step is to set the background to black and display, or store, the reconstructed images. The system described here presents a solution for storage, management and distribution of a unique large image dataset, namely VH color images, by efficient two step lossless compression, content based retrieval and interactive fast transmission and viewing. Such a system as described here provides innovative concepts for designing systems for other large databases by using optimization criteria for specific databases