VisionAI

Inspiration

Our project was born out of a desire to leverage technology for the betterment of the visually impaired community. We were inspired by the idea of using object detection to provide real-time audio feedback, empowering individuals with visual impairments to navigate their environment more independently.

What it does

Our project is a real-time object detection system integrated with a Raspberry Pi. It captures a live video feed using the Pi camera, processes it through a MobileNet SSD model, and instantly converts detected objects into spoken words using the pyttsx3 text-to-speech library. This system serves as a visual assistant, enhancing the daily lives of visually impaired individuals by providing auditory descriptions of their surroundings.

How we built it

We built this project by combining the power of open-source technologies. We used MobileNet SSD for object detection due to its speed and efficiency, which is crucial for real-time applications on a Raspberry Pi. The pyttsx3 library was employed to generate natural-sounding audio responses. Our development process involved setting up the Raspberry Pi, integrating the camera, and optimising the object detection model for low-latency audio output.

Challenges we ran into

During development, we faced challenges related to optimising the object detection process for real-time usage on the Raspberry Pi's hardware constraints. We also encountered issues with ensuring a seamless and responsive audio output system. However, through experimentation and problem-solving, we overcome these hurdles.

Accomplishments that we're proud of

We take pride in achieving a seamless integration of object detection and text-to-speech on a Raspberry Pi, resulting in a practical solution for visually impaired users. Additionally, we successfully optimised the system to operate in real-time, offering quick and accurate object recognition.

What we learned

This project provided valuable insights into computer vision, edge computing, and accessibility technology. We honed our skills in optimising machine learning models for resource-constrained devices, and we gained a deep appreciation for the importance of inclusive design in technology.

What's next

In our continuous efforts to enhance the functionality and user experience of our project, we have exciting plans on the horizon. One key aspect of our future development is the integration of additional sensors, such as infrared sensors for accurate distance measurement. This will empower our system to provide not only object recognition but also spatial awareness, enabling users to navigate their surroundings with even greater confidence and safety. Our aim is to create a comprehensive assistive technology solution that combines real-time object detection, spatial awareness, and natural language interaction to enrich the lives of visually impaired individuals.

Built With

• mobilenet-v3
• open-cv
• python
• pyttsx3
• raspberry-pi