Related Experiment Video
Updated: Apr 30, 2026
11:18
Closed-loop Neuro-robotic Experiments to Test Computational Properties of Neuronal Networks
Published on: March 2, 2015
10.3K
Motion sensitive network for action recognition in control and decision-making of autonomous systems
Jialiang Gu1, Yang Yi1, Qiang Li1
1Computer Science and Engineering, Sun Yat-sen University, Guangdong, China.
Frontiers in Neuroscience
|April 9, 2024
Summary
The Motion Sensitive Network (MSN) enhances video action recognition by effectively extracting motion information using novel spatial-temporal modules. This artificial neural network approach improves accuracy and shows potential for autonomous systems.
Area of Science:
- Artificial Intelligence
- Computer Vision
Background:
- Spatial-temporal modeling is vital for AI-driven action recognition in videos.
- Extracting motion information is challenging due to appearance changes and varied motion frequencies.
Purpose of the Study:
- To introduce the Motion Sensitive Network (MSN) for robust action recognition.
- To address limitations in extracting motion information from videos.
Main Methods:
- Developed the Spatial-Temporal Pyramid Motion Extraction (STP-ME) module for multi-scale temporal analysis.
- Introduced the Variable Scale Motion Excitation (DS-ME) module using differential models for flexible motion capture.
- Employed multi-scale deformable convolutions to adjust motion scales before temporal differencing.
Main Results:
- Achieved accuracy improvements of 1.1% to 2.2% on benchmark datasets.
- Reached a maximum performance of 89.90% compared to state-of-the-art methods.
- Ablation studies showed performance gains of 2% to 5.3%.
Conclusions:
- The Motion Sensitive Network (MSN) offers an effective framework for action recognition.
- MSN integrates artificial neural networks (ANNs) with autonomous system control concepts.
- Demonstrated significant potential for ANNs in challenging autonomous system applications.
