Abstract
In this paper, we mainly focus on automatic detection and segmentation of cell nuclei in histopathology images. Though some methods have been presented to solve these issues, there is scope for efficiency and performance to improve. We propose an end-to-end trainable convolutional neural network, which can learn the object-level and pixel-level information of image patches. In this way, the output feature map could be applied in nuclei detection and segmentation tasks concurrently. Then the weighted patch aggregation and refinement methods are utilized to achieve the final segmentation result. The experiments on the standard public dataset demonstrate that that our method achieves a good performance on nuclei detection and segmentation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have