Mouse Brain Cell Segmentation in Fluorescence Microscopy Images

Neurons exhibit remarkable morphological complexity, even within mouse brain images. These images provide critical insight into understanding neurological disorders such as Down syndrome and autism in mice, which serve as valuable models for human neurological conditions. By studying the fundamental units of neurons and nervous systems, researchers can uncover the genetic and neurobiological mechanisms underlying these disorders, offering potential pathways for therapeutic breakthroughs. Despite its significance, neuron morphology analysis remains one of the most challenging computational tasks in neuroscience due to the complexity and scale of the data. In this paper, we tackle this challenge by collecting high-resolution mouse brain images and systematically evaluating their analysis using six models, three transformer-based approaches and three non-transformer-based approaches. Our collected dataset provides an essential resource for biological researchers, facilitating deeper insights into the mechanisms of neurological disorders. The experimental results provide a detailed evaluation of these models’ performance and present a significant contribution to computational neuroscience community.