Consider an N x N image in which each pixel value is represented by k bits. By selecting a single bit from the same position in the binary representation of each pixel, an N x N binary image called a bit plane can be formed . For example, we can select the most significant bit of each pixel value to generate an N x N binary image representing the most significant bit plane. Repeating this process for the other bit positions, the original image can be decomposed into a set of k, N x N bit planes (numbered 0 for the least significant bit (LSB) plane through k â 1 for the most significant bit (MSB) plane). The motivation for this decomposition is that each bit plane can then be encoded efficiently using a lossless binary compression technique. Furthermore, in certain applications, the user may desire a low bit rate approximation to the original image before making the decision to proceed to a lossless mode. Since the more significant bit planes generally contain major structural information and are highly compressible, progressively reconstructing an image using the bit planes can be a viable technique for this purpose. This technique of progressive transmission is discussed in more detail in Chapter 14: Hierarchical Coding under Lossy Compression Techniques.
Online access to SPIE eBooks is limited to subscribing institutions.