Biomedical Lung Image Segmentation Using CNN

A variety of DNN is applied in segmentation of medical image and has achieved acceptable performance. Unlike natural pictures, medical pictures of a similar imaging modality are characterized by a similar pattern that indicates same traditional organs or tissues found at similar positions within the pictures. Thus, we tend to attempt to incorporate the previous information of medical pictures for correct segmentation into the structure of neural networks. Supporting this idea, we would like to propose a project paper on deep network known as Segmentation and analysis of Lung organ using CNN. The segmentation operate is analyzed. Experiments validate our information assumption concerning the incorporation of previous information into the convolution kernels and to show that we have attained an affordable segmentation and a satisfactory accuracy.