Task-Related Feature Enhancement Network for Neuronal Morphology Classification


Chunli Sun (University of Science and Technology of China), Feng Zhao (University of Science and Technology of China)
The 35th British Machine Vision Conference
PDFPosterVideo (Right click to download)

Abstract

Analyzing the morphology of pyramidal cells (PCs) is essential to understanding brain activity and disease mechanisms. Existing methods describe the morphology of PCs in a task-agnostic manner, leading to insufficient representations for various analysis tasks. This paper presents a Task-related Feature Enhancement Network (TFENet) to discern subtle morphological differences and identify the PCs in a task-related manner. The TFENet first extracts task-common features via a shared backbone across tasks and then generates task-specific features for each task individually. The task-specific features are refined using a Region Feature Enhancement Module (RFEM) based on the morphology-aware regions. Furthermore, a Global Category-guided Fusion Module (GCFM) adaptively combines the task-specific and task-common features, yielding a distinctive morphology descriptor. Extensive experiments demonstrate the effectiveness of our method, achieving 90.34% and 74.15% accuracy for the species and brain region analysis tasks, respectively, outperforming the task-agnostic methods.

Citation

@inproceedings{Sun_2024_BMVC,
author    = {Chunli Sun and Feng Zhao},
title     = {Task-Related Feature Enhancement Network for Neuronal Morphology Classification},
booktitle = {35th British Machine Vision Conference 2024, {BMVC} 2024, Glasgow, UK, November 25-28, 2024},
publisher = {BMVA},
year      = {2024},
url       = {https://papers.bmvc2024.org/0745.pdf}
}


Copyright © 2024 The British Machine Vision Association and Society for Pattern Recognition
The British Machine Vision Conference is organised by The British Machine Vision Association and Society for Pattern Recognition. The Association is a Company limited by guarantee, No.2543446, and a non-profit-making body, registered in England and Wales as Charity No.1002307 (Registered Office: Dept. of Computer Science, Durham University, South Road, Durham, DH1 3LE, UK).

Imprint | Data Protection