Neurons Detection Employing a Deep Convolution Neural Network

Neurological disorders, including neurodegenerative diseases such as Alzheimer's and brain tumors, are a leading cause of death and disability across the world. Accurate image segmentation of these cells with the help of computer vision could lead to new and effective drug discoveries to treat the millions of people with these disorders. In this paper, UNet++ (Resnet34) produces more accurate cell segmentation in challenging images such as high-resolution images and small cells. The results illustrate improvements to the broad UNet architecture, which is a robust architecture for segmenting medical image. The experiments demonstrate that UNet+ + with a supervised encoder-decoder structure achieves evaluation metrics with an average IoU of 58% and pixel accuracy of 96%. This method allows to detect the large numbers of medical cell in a short time and can be applied to automated cell imaging and analysis system.