Chapter 5:
Lossless Compression of Digital Mammographic Images
Editor(s): Jasjit S. Suri S. Vinitha Sree Kwan-Hoong Ng Rangaraj M. Rangayyan
Author(s): Mulemajalu, Ravikumar Koliwad, Shivaprakash
Published: 2012
DOI: 10.1117/3.899757.ch5

5.1 Introduction

An anonymous quote reads, "An image speaks more than ten thousand words." For mammographic images, this quote could be rephrased to read, "Images save more than ten thousand lives," as using mammography screening to detect cancer is effective in reducing breast cancer mortality rates by 30 to 70%. Breast cancer is the most common form of cancer in the human female, affecting an average of 1 in 11 women at some phase of their lives in the Western world. Currently, x-ray mammography is the clinical gold standard for the detection of breast cancer.

Mammographic images are large in size, high in resolution, and require voluminous storage space and transmission bandwidth. Sophisticated compression methods are required to reduce representative data while preserving clinical information. This chapter is an attempt to review various lossless compression algorithmic techniques, some of which have been exclusively developed for mammographic images.

The chapter is organized as follows: Section 5.2 provides an introduction to digital mammography and describes the importance of lossless compression in preserving clinical information while reducing the amount of data. Section 5.3 covers the storage of mammographic images and other related constraints. Section 5.4 gives an explanation of data compression and encoding and decoding. The basics of digital imaging, their storage, and measures used for performance comparison of lossless image compression are described in Section 5.5. Information theory basics and the application of these basics to lossless image compression are briefly discussed in Section 5.6. The basic definition of redundancy, types of redundancies, and their importance in lossless image compression are explained. Sections 5.7 and 5.8 deal with two prominent techniques used to remove (by substantial amounts) redundancy in digital images: spatial prediction and transform techniques. Lossless JPEG, JPEG-LS, and JPEG 2000 standard schemes that are incorporated for Digital Imaging and Communications in Medicine (DICOM) are also reviewed. Section 5.9 provides the performance comparisons of these compression schemes. Section 5.10 briefly reviews various methods used for the lossless compression of mammographic images. Section 5.11 contains descriptions of mammographic image databases. Section 5.12 lists analytical observations of the compression ratios of the reviewed techniques. Finally, a summary of the chapter is given in Section 5.13.

Online access to SPIE eBooks is limited to subscribing institutions.

Back to Top