Related Experiment Video
Updated: Jan 2, 2026
09:29
A Standardized Obstacle Course for Assessment of Visual Function in Ultra Low Vision and Artificial Vision
Published on: February 11, 2014
13.4K
Object Identification and Safe Route Recommendation Based on Human Flow for the Visually Impaired
Yusuke Kajiwara1, Haruhiko Kimura1
1Department of Production Systems Engineering and Sciences, Komatsu University, Awazu Campus, Komatsu 923-0921, Japan.
Sensors (Basel, Switzerland)
|December 11, 2019
Summary
This study introduces "Follow me!", an app using smartphone cameras to analyze pedestrian gait for safer navigation for visually impaired individuals. The system achieved 92% accuracy in identifying pedestrians and obstacles, guiding users safely 100% of the time.
Area of Science:
- Computer Science
- Human-Computer Interaction
- Robotics
Background:
- Visually impaired individuals face significant mobility challenges, compounded by limitations in current navigation systems.
- Existing obstacle detection methods struggle in high-traffic areas due to occlusion, reducing accuracy.
Purpose of the Study:
- To develop an accurate and reliable navigation system for the visually impaired.
- To overcome the limitations of traditional obstacle detection in crowded environments.
Main Methods:
- Developed the "Follow me!" application utilizing machine learning on smartphone monocular camera images.
- Analyzed pedestrian gait and walking routes to predict safe navigation paths.
- Trained the system to identify same-direction pedestrians, oncoming pedestrians, and steps.
Main Results:
- Achieved an average accuracy of 0.92 in identifying pedestrians and environmental features based on gait and route analysis.
- The recommended safe routes demonstrated 100% accuracy in guiding visually impaired users.
- Successfully enabled users to avoid temporary obstacles like construction and signage.
Conclusions:
- The "Follow me!" application offers a novel and effective solution for visually impaired navigation.
- Machine learning analysis of pedestrian dynamics significantly enhances navigation safety in complex environments.
- This technology has the potential to improve the independence and mobility of visually impaired individuals.
