仿果蝇复眼视神经的快速目标追踪模型代码

基于果蝇神经重叠型复眼结构特征、复眼视觉目标背景分离和追踪的并行神经计算的机制，通过对果蝇运动视觉神经系统的行为和生理研究、目标定位与追踪相关的视觉功能机制解析，建立果蝇进行高速避碰与追踪的分割-识别-追踪流程框架。实现运动背景下基于先验条件下的目标快速识别以及基于应激神经通路的追踪与避碰方法，构建出一套完整可用的视觉处理系统。Based on the structural characteristics of the overlapping compound eyes of Drosophila and the mechanism of parallel neural computation for visual target-background separation and tracking of the compound eye, this dataset is built through behavioral and physiological studies on the motion visual nervous system of Drosophila and the analysis of visual functional mechanisms related to target localization and tracking. A segmentation-recognition-tracking workflow framework for high-speed obstacle avoidance and tracking in Drosophila is established, realizing rapid target recognition under prior conditions in moving backgrounds and enabling tracking and obstacle avoidance methods based on stress-related neural pathways, thus constructing a complete and practical visual processing system.

搜集汇总

背景与挑战

背景概述

该数据集提供了基于果蝇复眼视觉神经网络的快速目标追踪模型代码，包括分割、识别和跟踪过程的框架实现。数据集包含14个MATLAB代码文件，总大小为33.08 KB，由中国科学院自动化研究所的李宁于2023年12月发布，适用于目标追踪和仿生视觉研究。