Seeing Aid for Visually Impedpeded Pedestrians Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Visually impaired pedestrians face challenges in daily life due to inability to recognize objects, perceive dangers, and read text, with existing assistive technologies being limited in functionality and reliance on external assistance.
Innovation Solution
A seeing aid system comprising a camera, processing unit, and audio output unit that uses machine learning to provide verbal descriptions of the environment, allowing users to learn object names through speech input, and includes features like infrared detection, ambisonic audio, and vibration units for enhanced safety and functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If guide dogs are used to assist blind people, then basic navigation and safety guidance are improved, but the ability to provide detailed information about objects, people, and environment remains limited
Solution Approach 1:
The patent replaces the mechanical/biological system of guide dogs with an electronic system comprising cameras, image processing units, and audio output devices. This substitution enables comprehensive environmental analysis and detailed information delivery about objects, people, and surroundings that a guide dog cannot provide.
Solution Approach 2:
The patent introduces an intermediary system (image processing device) between the blind person and the environment. This intermediary captures visual information through cameras, processes it to identify objects and features, and delivers detailed information through audio output, thereby bridging the information gap that guide dogs cannot fill.
2Ease of operation
If blind people call assistants for help, then they can obtain external assistance, but they remain dependent on outside help and cannot navigate independently
Solution Approach 1:
The patent enables blind people to independently obtain environmental information through self-contained devices worn on the body. The system autonomously captures images, processes them to identify objects and features, and delivers information without requiring external human assistance, thereby achieving self-service and reducing dependency on others.
3Loss of information
If traditional camera systems are used to capture images, then visual information is recorded, but the system cannot automatically identify and provide meaningful information about objects
Solution Approach 1:
The patent incorporates pre-trained machine learning models and object recognition algorithms in the image processing device. These preliminary computational structures enable automatic identification of objects, people, and environmental features directly from captured images, converting raw visual data into meaningful information without requiring complex real-time analysis.
Solution Approach 2:
The image processing device is designed with multi-functional capabilities including object recognition, text detection, scene understanding, and feature identification. This universal processing system handles various types of visual information through integrated algorithms, providing comprehensive environmental awareness without requiring multiple separate specialized devices.
Data Source
Figure 1a~1b
Figure 2a~2b
Figure 3
AI summary
The present invention provides a seeing aid for a visually impaired pedestrian, wherein the seeing aid comprises: a camera for recording an image of an environment of the pedestrian, a processing unit for processing the image to obtain verbal information about the environment, and an output unit for outputting the verbal information to notify the pedestrian.