期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2021
卷号：9
期号：7
页码：8006-8013
DOI：10.15680/IJIRCCE.2021.0907027
语种：English
出版社：S&S Publications
摘要：Cancer is very dangerous disease that causes millions of death throughout the world. Various parts of human body can be affected with cancer cells. Liver cancer becomes very dangerous and it was found in most of the cases in recent days. It is very difficult to find liver cancer in initial stage. The recognition and segmentation of liver cancer from medical images is a major task. In this project, we propose a novel liver cancer image segmentation method based on Convolutional Neural Network (CNN). The main structure of a convolutional neural network adopts U-NET (‘U’ shaped network). U-NET is image segmentation. The network is based on the fully convolutional network and its architecture was modified and extended to work with fewer training images and to yield more precise segmentations. In order to capture objects of different scales in the deep features, a group of cascaded dilated convolution is inserted at the bottom of U-NET, which has different dilation rates. Furthermore, to better optimize the network at different scales, an auxiliary loss function is proposed to be integrated in the cascaded dilated convolution. We have proposed an effective and efficient hybrid architecture for extraction of liver cancer. Hence, liver diseases can be diagnosed using this technique and also be classified using the same using advanced computational techniques and large dataset. The system can match the results of a liver cancer thus improving the quality standards in the area of medicine and research.
