Junhao Yang¹, Chunxiao Chen^1,*, Qingyang Zang¹, Jianfei Li¹

Molecular & Cellular Biomechanics, Vol.15, No.4, pp. 203-214, 2018, DOI:10.32604/mcb.2018.03824

Abstract Pathological slide is increasingly applied in the diagnosis of breast tumors despite the issues of large amount of data, slow viewing and high subjectivity. To overcome these problems, a micrograph recognition method based on convolutional neural network is proposed for pathological slide of breast tumor. Combined with multi-channel threshold and watershed segmentation, a sample database including single cell, adhesive cell and invalid cell was established. Then, the convolution neural network with six layers is constructed, which has ability to classify the stained breast tumor cells with accuracy of more than 90%, and evaluate the proliferation level with relative error of… More >

Wei Fang^1,2, Feihong Zhang^1,*, Victor S. Sheng³, Yewen Ding¹

CMC-Computers, Materials & Continua, Vol.57, No.1, pp. 167-178, 2018, DOI:10.32604/cmc.2018.02356

Abstract Image recognition has always been a hot research topic in the scientific community and industry. The emergence of convolutional neural networks(CNN) has made this technology turned into research focus on the field of computer vision, especially in image recognition. But it makes the recognition result largely dependent on the number and quality of training samples. Recently, DCGAN has become a frontier method for generating images, sounds, and videos. In this paper, DCGAN is used to generate sample that is difficult to collect and proposed an efficient design method of generating model. We combine DCGAN with CNN for the second time.… More >