Accession Number:

ADA300271

Title:

Digital Mammographic Image Compression.

Descriptive Note:

Annual rept. 15 Jul 94-14 Jul 95,

Corporate Author:

CALIFORNIA UNIV SAN FRANCISCO

Personal Author(s):

Report Date:

1995-07-15

Pagination or Media Count:

23.0

Abstract:

This research investigated both lossless and lossy compression methods for digital mammography. A structure lossless compression algorithm was developed for digital mammograms. The algorithm utilizes the unique shape and image characteristics of mammograms. The algorithm first segments a breast image from its background. An averaging predict coding was then used to compress the segmented image. The average bit rate compression ratio achieved is about 3.4 for 100 mammograms. The results also show that high resolution images can be compressed more than low resolution images. In the lossy compression, a wavelet compression scheme was developed. The preliminary results show that the wavelet transform is able to concentrate pixel energy into a very small area. This characteristic is promising for obtaining high quality compressed images at reasonable compression ratios.

Subject Categories:

  • Medicine and Medical Research
  • Cybernetics

Distribution Statement:

APPROVED FOR PUBLIC RELEASE