Related Experiment Video
Updated: Sep 11, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
635
Leveraging assistive technology for visually impaired people through optimal deep transfer learning based object
Mahir Mohammed Sharif Adam1, Nojood O Aljehane2, Mohammed Yahya Alzahrani3
1Department of Computer and Self Development, Preparatory Year Deanship, Prince Sattam bin Abdulaziz University, AlKharj, Saudi Arabia. m.adam@psau.edu.sa.
Scientific Reports
|August 17, 2025
Summary
This study introduces an enhanced assistive technology for blind individuals, improving object detection accuracy using deep learning and a novel optimization algorithm. The developed technique achieves 99.25% accuracy, significantly aiding visually impaired people.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Assistive Technology
Background:
- Visual impairment significantly impacts cognitive and psychological well-being, necessitating effective assistive technologies.
- Existing research often prioritizes mobility and navigation, with less focus on aesthetic aspects of assistive tools for the visually impaired.
- Object detection is vital for computer vision applications, with deep learning techniques driving recent advancements.
Purpose of the Study:
- To develop an effective object detection model for visually impaired individuals using advanced deep learning techniques.
- To enhance the quality of life for people who are blind through improved assistive technology.
- To address the need for better object recognition capabilities in assistive tools for the visually impaired.
Main Methods:
- Proposed an enhanced assistive Technology for Blind People through Object Detection Using a Hiking optimization algorithm (EATBP-ODHOA).
- Employed adaptive bilateral filtering (ABF) for image pre-processing to reduce noise.
- Utilized Faster R-CNN with ResNet and DenseNet-201 fusion models for feature extraction and bidirectional gated recurrent unit (Bi-GRU) for classification.
- Optimized model parameters using the Hiking Optimisation Algorithm (HOA).
Main Results:
- The EATBP-ODHOA technique achieved a superior accuracy of 99.25% in object detection.
- Experimental results on an indoor object detection dataset demonstrated the model's effectiveness.
- The proposed method outperformed existing approaches in accuracy for assisting visually impaired individuals.
Conclusions:
- The EATBP-ODHOA technique offers a significant advancement in assistive technology for the visually impaired.
- The integration of deep learning, fusion models, and a novel optimization algorithm enhances object detection capabilities.
- This technology has the potential to greatly improve the daily lives and independence of people with visual impairments.
