This paper deals with classification of sarcomas using textural features. We assume that sarcomas have single texture and the same scale is used for photographing sarcoma's images. 80 sorts of sarcomas are selected under those assumptions. First, we make database of templates of sarcomas using textural features. The intensity histogram, Fourier power spectrum, run-length matrix, fractal dimension and co-occurrence matrix are used as textural features. Second, sarcomas are collated by using template matching between unknown samples and the templates in the database. The success rate reached 92%. The result of this experiment shows the adequateness of proposed system.
|Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings
|Published - 2000 12月 1
|22nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society - Chicago, IL, United States
継続期間: 2000 7月 23 → 2000 7月 28
ASJC Scopus subject areas
- コンピュータ ビジョンおよびパターン認識