Abstract
Cancer is the most cause of death in Japan, and patients suffering with and who die of cancer are increasing every year, while the number of pathologists is almost constant. In this paper, an automatic cancer detection method that combines multiple features to support pathologists is presented. We use three features, Higher-order Local Auto- Correlation (HLAC) feature, wavelet feature, delaunay feature. At first, features are calculated from gastric lymph node image. Then we combine features, and distinguish cancer and non-cancer area using Support Vector Machine (SVM). HLAC, wavelet and delaunay feature is shape, frequency, and nuclear-position geometrical characteristics respectively. Combination of all three features is the best accuracy rate, and its sensitivity and specificity are 94.6% and 84.9% respectively. Accuracy rate that combines more two features is better than only one feature.
