Improving of colon cancer cells detection based on Haralick's features on segmented histopathological images
Ahmad Chaddad, Camel Tanougast, A. Dandache, Ali Al Houseini, Ahmed Bouridane · 2011
Image analysis in cancer pathology applications has evolved considerably in the last years [1]. The areas concerned were particularly those in which the diagnosis was based on the medical image processing and analysis. Few studies have successfully investigated the automatic classification of colonic pathology images if they contain healthy cells or cancerous cells. The objective of this work is the multispectral images classification of healthy and cancerous cells in order to accelerate the operations of classification between different types of cancerous cells. Our detection approach was derived from the "Snake" method but using a progressive division of the dimensions of the image to achieve faster segmentation. The time consumed during segmentation was decreased to more than 50%. We extract several Haralick's coefficients to detect the type of cells were made segmentation are applied to the multispectral image. The experimental results obtained on several multispectral images show that the method is efficient for the classification of cancer cells of type Carcinoma (Ca), Intraepithelial Neoplasia (IN) and Benign Hyperplasia (BH).