Rich feature hierarchies for cell detecting under phase contrast microscopy images

R-CNN (region-convolutional neural network) has recently achieved very outstanding results in variety of visual detecting fields, and its function of object-proposal-generation can achieve effective training models by using as small samples as possible in the field of machine learning. In this paper, a modified R-CNN is proposed and applied to detect cells under phase contrast microscopy images by adopting multiple object-proposal-generations instead of a single one to extract candidate regions. The results show that the proposed method can obtain better performance than the traditional method by using a single object-proposal-generation.