Effects of adaptive speckle filtering on the diagnostic accuracy of ultrasound in vivo tissue characterization
✍ Scribed by N. Suhm; I. Zuna; M. Fein; G. Weisser
- Publisher
- Elsevier Science
- Year
- 1995
- Tongue
- English
- Weight
- 543 KB
- Volume
- 2
- Category
- Article
- ISSN
- 0929-8266
No coin nor oath required. For personal study only.
✦ Synopsis
The purpose of this study was to investigate whether there is an influence of adaptive speckle filtering on ultrasonic tissue characterization. Methods: We have implemented two adaptive two-dimensional speckle filters, the multivariate comparator and the local statistics filter. A tissue characterization system was used to classify ultrasound images showing diffuse liver disease. We compared the classification results achieved before and after speckle filtering. Results: The image texture was completely smoothed out and the classification results got worse when the local statistics filter had been used for adaptive speckle filtering. On the other hand, the image texture was preserved to a certain degree and the classification results were equally good or even slightly improved when the multivariate comparator had been used. In addition, the performance of both filters was demonstrated when they were applied to ultrasound images showing focal liver disease. Conclusion: The two adaptive speckle filters we have investigated differ in their effect on the ~ texture. The classification results of ultrasonic tissue characterization are reversely influenced by the two adaptive speckle filters. This can be explained with the presence of relevant information in the speckle pattern of homof~-neous tissue in ultrasound images.