Proceedings of the KOSOMBE Conference (대한의용생체공학회:학술대회논문집)
- Volume 1992 Issue 11
- /
- Pages.26-28
- /
- 1992
The Ultrasound Image Diagnosis using Statistical Characteristics and Neural Network
통계적 특성과 신경망을 이용한 초음파 화상진단
- Hong, Jeong-Woo (Dept. of Electronics Eng., Hanyang University) ;
- Kim, Sun-Il (College of Medicine, Hanyang University) ;
- Lee, Doo-Soo (Dept. of Electronics Eng., Hanyang University)
- Published : 1992.11.13
Abstract
Texture analysis, one of the mage processing techniques, using statistical characteristics is applied to the ultrasound images, which are then classified into each types through neural network. This is a method to be used to diagnose ultrasound images automatically and objectively. First tone kinds of texture analysis techniques proposed already are used to classify ultrasound images and compared in terms of classification rate, and then a new technique if proposed which is invariant to multiplicative gain changes and image resolution.
Keywords