Compression of vector confocal and video images of microscopic objects by quantization and image resampling

Abstract

A new technique for lossy compression of grayscale images obtained from video and confocal microscopy is presented. Video microscopy allows the observation of moving microscopic specimens, while confocal microscopy can resolve two-dimensional sections of three dimensional microscopic objects. In both cases a single specimen can yield a series of digital images, each one being slightly similar to the others in that series. Each image in a series is called a frame. This paper will show that image compression using vector quantization (VQ) exploits the interframe redundancy inherent in these series of images. This is achieved without the need for any special interframe processing used in popular transform-based techniques like the MPEG.