Morphometric pattern analysis of basal cell nuclei for oral cancer screening
This work presents a quantitative approach fordiscrimination of Oral Submucous Fibrosis (OSF) to NormalOral Mucosa (NOM) in respect to size and shape properties of the basal layer, first layer in epithelium. Practically, basal cellsform the proliferative compartment of the epithelium, andtherefore changes in the morphometry of basal cells may haveserious implications on future cell behavior, includingmalignant transformation according to onco-pathologistsvision. In view of this, the changes in shape and size of thenuclei in the basal cell layer of the oral epithelium have beenstudied here by developing an automated image analyzer.Geometric, Zernike moments and transformation basedfeatures are extracted for morphometric pattern analysis of thenuclei. These features are statistically analyzed along with 3Dvisualization in order to discriminate the groups. Resultsshowed increase in the dimensions (area and perimeter) andshape parameters of the nuclei from normal mucosa to OSFwith dysplasia. Finally, pattern analyzer is employed usingBayesian approach and error back-propagation neuralnetwork. The performance is evaluated by partitioning thewhole data set into various combinations of training-testingsubsets, finally which converge to overall accuracies 97.02%for neural network and 97.93% for Bayesian classifierrespectively.
Pratik Shah Muthu Rama KrishnanM Chandan Chakraborty Ajoy Kumar Ray
School of Medical Science and Technology, Indian Institute of Technology-KharagpurKharagpur, West Be Chandan Chakraborty Bengal Engineering and Science University Kolkata, India
国际会议
成都
英文
1-4
2010-06-18（万方平台首次上网日期，不代表论文的发表时间）