Vessel Detection for Diagnosis of Tumor Invasion with Differently Stained Pathology Images

抄録

In the medical filed, detection and classification of cancer cells play important roles in analysing cells and developing new medicines. Performing them have drawbacks such as time-consuming and low accuracy. In this study, it is focus to solve these two by marking out the suspicious regions of vessels in the pathology images with HE and EVG staining systems. Next, collecting pixel’s data using Slide window method and splitting data are performed to get the training and testing data which are used as an input data in Neural Network later. Neural Network model with Multilayer Perceptron is used for training and testing the data.