Computer-assisted Colonoscopy Diagnostic Imaging Technology

High incidence of cancer cases in Asia has triggered a need to develop an effective automated screening method for early detection. Currently computer-based colonoscopic diagnosis for suspected malignancies has not been employed. Therefore the objective of this dissertation is to develop a new technique for automated analysis of the colonoscopic images based on multi-information.Aiming at color colonoscopic images, a new algorithm for segmenting color colonscopic images by fusing multiple information, such as color, brightness, spatial distance and texture and then by stochastic clustering has been presented in this thesis. It makes the fractal dimension (FD) as the measurement for texture feature in images and employs a stochastic clustering algorithm that uses pairwise similarity of elements. The clustering algorithm, based on a new graph theoretical algorithm for the sampling of cuts in graphs, can obtain the optimal number of clusters automatically. The complexity of the algorithm is simpler, and its stochastic nature makes it robust against noise.A multilayer perceptron with backpropagation learning is selected for classifying colonoscopic images into normal and potentially abnormal categories. The results show that the neural network is more appropriate for the classification of colon status.The new algorithm in this paper has been tested by a number of colonoscopic images. Initial results suggest the feasibility of the proposed method for colorectal cancer screening. The proposed new computer-assisted colonoscopic analysis method may prove to be beneficial for early detection of cancer and for better patient management.