Improved faster RCNN for white blood cells detection in blood smear image
Cheng Sun, Suhua Yang, Jiang Shaofeng · 2019
The inspection result of the white blood cell count is actually a critical part of medical diagnose. In the past five years, with the application of deep learning technique in the field of recognition of white blood cell, the classification accuracy of the white blood cell has been improved dramatically, however the extraction accuracy is still low, which limits the improvement of the total accuracy. In this paper, to solve the above problems, we proposed an algorithm of white blood cell recognition based on improved Faster RCNN. We improved RPN network by optimized the parameters of the original RPN network to make it have the ability to detect the small target. The results showed that the F1-Score improved compared with original Faster RCNN on self-built data set. The results show our method can extract the adhesion cells.