NRTR: Neuron Reconstruction with Transformer from 3D Optical Microscopy Images
This addresses the problem of time-consuming and inefficient neuron reconstruction for neuroscience, offering a simpler end-to-end approach, though it appears incremental as it builds on existing deep learning methods.
The paper tackled neuron reconstruction from 3D optical microscopy images by proposing NRTR, an end-to-end method that views it as a set-prediction problem, achieving excellent results on BigNeuron and VISoR-40 datasets and outperforming baselines.
The neuron reconstruction from raw Optical Microscopy (OM) image stacks is the basis of neuroscience. Manual annotation and semi-automatic neuron tracing algorithms are time-consuming and inefficient. Existing deep learning neuron reconstruction methods, although demonstrating exemplary performance, greatly demand complex rule-based components. Therefore, a crucial challenge is designing an end-to-end neuron reconstruction method that makes the overall framework simpler and model training easier. We propose a Neuron Reconstruction Transformer (NRTR) that, discarding the complex rule-based components, views neuron reconstruction as a direct set-prediction problem. To the best of our knowledge, NRTR is the first image-to-set deep learning model for end-to-end neuron reconstruction. In experiments using the BigNeuron and VISoR-40 datasets, NRTR achieves excellent neuron reconstruction results for comprehensive benchmarks and outperforms competitive baselines. Results of extensive experiments indicate that NRTR is effective at showing that neuron reconstruction is viewed as a set-prediction problem, which makes end-to-end model training available.