Combining Deep Learning and Active Contours Opens The Way to Robust, Automated Analysis of Brain Cytoarchitectonics

Abstract

Deep learning has thoroughly changed the field of image analysis yielding impressive results whenever enough annotated data can be gathered. While partial annotation can be very fast, manual segmentation of 3D biological structures is tedious and error-prone. Additionally, high-level shape concepts such as topology or boundary smoothness are hard if not impossible to encode in Feedforward Neural Networks. Here we present a modular strategy for the accurate segmentation of neural cell bodies from light-sheet microscopy combining mixed-scale convolutional neural networks and topology-preserving geometric deformable models. We show that the network can be trained efficiently from simple cell centroid annotations, and that the final segmentation provides accurate cell detection and smooth segmentations that do not introduce further cell splitting or merging.

Mapping Mesoscale Axonal Projections in the Mouse Brain Using A 3D Convolutional Network

“…The automated identification and segmentation of axons from 3D images should circumvent these limitations. Recent application of deep convolutional neural networks (DCNNs) and Markov random fields to biomedical imaging have made excellent progress at segmenting grayscale CT and MRI volumes for medical applications (Alegro et al, 2017;Dong et al, 2018;Frasconi et al, 2014;Mathew et al, 2015;Thierbach et al, 2018). Other fluorescent imaging strategies including light-sheet, fMOST, and serial two-photon tomography have been combined with software like TeraVR, Vaa3D, Ilastik, and NeuroGPS-Tree to trace or otherwise reconstruct individual neurons (Peng et al, 2010;Quan et al, 2016;Wang et al, 2019;Winnubst et al, 2019;Zhou et al, 2018).…”