Inspiration
The idea for ODG-Vision was sparked by the desire to make daily life more accessible and safer for visually impaired individuals. Witnessing their struggles in navigating unfamiliar spaces and interacting socially inspired us to create a wearable device that could assist in these areas. With advances in AI, we saw an opportunity to develop a prototype that combines object detection, facial recognition, and voice-guided interaction into a single, intuitive wearable.
What it does
ODG-Vision is a compact wearable device that assists visually impaired users by: • Detecting Objects: Identifies and notifies users of nearby objects, enabling safe navigation. • Recognizing Faces: Uses facial recognition to identify familiar faces, like friends or family, helping users feel connected. • Voice Guidance: Provides real-time feedback through voice, guiding the user and confirming detected objects and faces.
How we built it
- Face Detection and Recognition: Using OpenCV, we employed LBPHFaceRecognizer and haarcascade classifiers to detect both frontal and profile faces. This allows the device to recognize users even from various angles.
- Voice Interaction: Python libraries like pyttsx3 for text-to-speech and speech_recognition enabled us to create a voice-guided experience. Users receive real-time audio cues for object and face recognition, making the interface hands-free and intuitive.
- Hardware Setup: The device is equipped with a camera, microphone, and speaker for capturing visual and audio input and delivering audio feedback. We implemented on-device processing to provide fast responses.
Challenges we ran into
• Multi-Angle Face Detection: Achieving consistent face recognition from different angles required the use of multiple face detection methods. • Real-Time Processing: Ensuring that recognition and feedback were timely and smooth demanded optimizations in the code for image capture and processing. • Ambient Noise and Voice Recognition: Accurate voice recognition was challenging in noisy environments. Implementing error handling and retry mechanisms helped mitigate this issue.
Accomplishments that we're proud of
• Successfully implemented a reliable multi-angle face detection system that supports both frontal and profile views. • Developed a real-time voice feedback system that is intuitive and effective for visually impaired users. • Created a user-friendly, accessible interface that allows users to navigate their surroundings independently and safely.
What we learned
Working on ODG-Vision taught us the importance of accessibility in technology and the specific challenges involved in building for assistive needs. We gained a deeper understanding of optimizing AI models for real-time applications, handling diverse environmental factors, and creating a streamlined experience for users with specific accessibility requirements.
What's next for ODG Vision
• Enhanced Object Detection: Expanding the object detection database to include more types of everyday items. • Improved Voice Recognition: Integrating advanced noise-canceling algorithms to improve voice recognition in loud environments. • Battery Optimization: Reducing power consumption to extend battery life for prolonged usage. • Mobile App Integration: Developing a mobile app that allows caregivers to track and monitor the device’s use, adding another layer of safety and connection.
Log in or sign up for Devpost to join the conversation.