Related Experiment Video
Updated: Jun 27, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
522
Smart traffic management of vehicles using faster R-CNN based deep learning method
1Great Lakes Institute of Management, Chennai, India. arindamphdthesis@gmail.com.
Scientific Reports
|May 6, 2024
Summary
This study introduces a Faster R-CNN deep learning method for accurate vehicle segmentation in smart traffic management. The approach enhances traffic density estimation and vehicle tracking, even in challenging conditions like occlusions and varying traffic density.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Transportation Engineering
Background:
- Smart traffic management is crucial for modernizing cities.
- Vehicle segmentation is key for traffic density estimation and speed monitoring.
- Challenges include occlusions, clutter, and varying traffic density.
Purpose of the Study:
- To investigate a Faster R-CNN deep learning method for vehicle segmentation.
- To improve the accuracy of vehicle segmentation in complex traffic scenarios.
Main Methods:
- Utilized a Faster R-CNN based deep learning framework.
- Employed adaptive background modeling to handle illumination and shadows.
- Incorporated extended topological active nets for refined segmentation and deformation.
- Optimized results through energy minimization and mesh deformation.
Main Results:
- Achieved higher segmentation accuracy compared to existing methods.
- Demonstrated the framework's superiority in experimental results.
- Successfully addressed issues of occlusion, clutter, and density variations.
Conclusions:
- The proposed Faster R-CNN framework offers a robust solution for vehicle segmentation in smart traffic management.
- The integration of topological active nets significantly enhances segmentation precision.
- This method provides a foundation for more effective intelligent transportation systems.
