Segmenting Neuronal Structure in 3D Optical Microscope Images Via Knowledge Distillation with Teacher-Student Network

Heng Wang,Donghao Zhang,Yang Song,Siqi Liu,Yue Wang,Dagan Feng,Hanchuan Peng,Weidong Cai

DOI: https://doi.org/10.1109/isbi.2019.8759326

2019-01-01

Abstract:Three-dimensional (3D) volumetric neural image segmentation is crucial to reconstructing accurate neuron structures. However, due to the structural complexity of neurons and the diverse imaging qualities of the microscopes, it is challenging to achieve both accuracy and efficiency. In this paper, we propose a teacher-student learning framework for fast neuron segmentation. The segmentation inference is performed using a light-weighted student network which benefits from knowledge distillation of a teacher network with a higher capacity. Evaluated on the Janelia dataset from the BigNeuron project, our proposed framework achieves competitive performance for segmentation accuracy while reducing the computational cost to facilitate large-scale processing.