DL segmentation for immunofluorescent imaging of mouse brains

Illustration (click to hide):

Project Description

In this project, we are looking at immunofluorescent staining of mouse brain cryosections for the study of cerebral cavernous malformations. A necessary task for analysing each image is marking the cerebellum, in effect differentiating this brain region from other neighbouring brain structures present in the image. A U-Net model with a ResNet-50 encoder was trained with previously generated manual annotations to automate this process. The resulting python-based tool carries out cerebellum segmentation, generating a predicted region of interest (ROI) from a supplied DAPI image.