Application of Grey Relational Analysis to Recognition of Liver Cancer in Biopsy Images
Shih-Ming Pan, Chia‐Hung Lin · 2012
For the diagnosis of liver cancer using a biopsy technique, pathologists' decisions are mainly based on the spatial and texture information of the biopsy images. However, the diagnostic accuracy strongly depends on the pathologist's knowledge and experience, that is, such diagnostic results are subjective. Hence, to make an objective and high accuracy diagnosis for the liver cancer in biopsy images, an efficiently software system is developed in this study. To well characterize the liver biopsy images, this study estimates their fractal dimensions as texture features to distinguish normal and cancerous liver tissue. Based on these fractal features, a grey relation analysis technique is applied to construct a pattern classifier as an objectively and efficiently recognition system. Experimental results show that the developed pattern classifier has good accuracy for the recognition of liver cancer in biopsy images.