# Structure-Aware Multi-Animal Pose Estimation for Space Model Organism Behavior Analysis

## TL;DR

This paper introduces a new deep learning method for accurately estimating the body poses of multiple small animals in space experiments, enabling detailed behavior analysis under microgravity.

## Contribution

A novel single-stage multi-animal pose estimation method with anatomical priors and structure-guided learning for robust performance in space conditions.

## Key findings

- The method achieves AP scores of 72.8%, 62.1%, and 67.1% on C. elegans, zebrafish, and Drosophila, surpassing state-of-the-art performance.
- The proposed approach supports end-to-end inference without separate detection or grouping steps, improving efficiency.
- It demonstrates robustness under occlusion and overlap, making it suitable for space biology experiments.

## Abstract

In space biology experiments, accurately estimating the body poses of small animals—such as Caenorhabditis elegans, zebrafish, and Drosophila—under complex conditions is essential for understanding how microgravity and other space-related factors affect behavior. This study proposes a deep learning-based multi-animal pose estimation method tailored to challenging scenarios involving dense populations, diverse postures, and frequent occlusions. By integrating anatomical priors, multi-scale visual features, and a structure-guided learning mechanism, the method demonstrates robust keypoint localization even under overlapping or partially occluded conditions. It supports end-to-end inference without the need for separate object detection or instance grouping steps, and achieves higher efficiency than traditional top–down and bottom–up approaches. The method was evaluated on the public SpaceAnimal dataset, which includes three types of space-experiment animals. Results show that it consistently outperforms existing baselines in pose estimation accuracy while maintaining a good balance between accuracy and efficiency. These findings highlight the method’s potential to enable automated, fine-grained behavior analysis in space life science research. As this study used only publicly available video data from previous space missions, no new animal experiments were conducted, and no ethical approval was required.

Multi-animal pose estimation is a critical technique for enabling fine-grained quantification of group animal behaviors, which holds significant scientific value for uncovering behavioral changes under space environmental factors such as microgravity and radiation. Currently, the China Space Station has conducted a series of space biology experiments involving typical model organisms, including Caenorhabditis elegans (C. elegans), zebrafish, and Drosophila. However, substantial differences in species types, body scales, and posture dynamics among these animals pose serious challenges to the generalization and robustness of traditional pose estimation methods. To address this, we propose a novel, flexible, and general single-stage multi-animal pose estimation method. The method constructs species-specific pose group representations based on anatomical priors, incorporates a multi-scale feature-sampling module to integrate shallow and deep visual cues, and employs a structure-guided learning mechanism to enhance keypoint localization robustness under occlusion and overlap. We evaluate our method on the SpaceAnimal dataset—the first public benchmark for pose estimation and tracking of model organisms in space—containing multi-species samples from both spaceflight and ground-based experiments. Our method achieves AP scores of 72.8%, 62.1%, and 67.1% on C. elegans, zebrafish, and Drosophila, respectively, surpassing the state-of-the-art performance. These findings demonstrate the effectiveness and robustness of the proposed method across species and imaging conditions, offering strong technical support for on-orbit behavior modeling and large-scale quantitative analysis.