Towards Computational Models and Applications of Insect Visual Systems for Motion Perception: A Review

Area of Science:

• Neuroscience
• Computer Science
• Robotics

Background:

• Motion perception is vital for insect survival, aiding in predator avoidance and foraging.
• Insect visual systems possess specialized motion detectors crucial for survival.
• Computational models of insect motion detection offer insights into biological systems and AI.

Purpose of the Study:

• To review computational motion perception models derived from insect visual systems.
• To explore the application of these bio-inspired models in robotics and artificial intelligence.
• To summarize methodologies for achieving direction and size selectivity in motion detection.

Main Methods:

• Review of computational models of insect motion perception.
• Analysis of looming-sensitive neuronal models (lobula giant movement detectors - LGMDs) in locusts.
• Examination of translation-sensitive direction-selective neurons (DSNs) in various insects.
• Investigation of small-target motion detectors (STMDs) in dragonflies and hoverflies.

Main Results:

• Identified key insect motion detection models: LGMDs, DSNs, and STMDs.
• Demonstrated the utility of these models in artificial intelligence and robotic applications.
• Summarized techniques for generating direction and size selectivity in motion perception.

Conclusions:

• Insect-derived motion perception models provide robust frameworks for artificial intelligence.
• Bio-inspired dynamic vision systems can be constructed using these biological mechanisms.
• Further research into systems integration and hardware realization of these models is warranted.