Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

GitHub - o40/seesay: Live image description solution using ESP32-CAM + Phone + Server

Jan 05, 2025 - github.com
The article describes a low-cost tool designed to assist the visually impaired by providing live descriptions of scenes captured by a camera. The author used an ESP32-CAM with built-in WiFi to capture images, which are then described using the gpt-4o-mini AI model and read back to the user via voice synthesis. The setup involves a cell phone with internet sharing, an HTTP server, and a power bank for the ESP32-CAM. While the proof-of-concept works, it has limitations such as the need for a web page to be open on a cell phone for descriptions, difficulty in mounting the camera, and lack of security.

The author tested the solution and found that while it works, the descriptions often include unnecessary details like weather and location. By refining the prompts, the descriptions improved but still required further enhancement. The project aims to provide a more affordable alternative to expensive existing products like Envision Glasses and OrCam MyEye. The author expresses interest in further development if suitable hardware with an open API becomes available, particularly for integrating the camera into glasses for better usability.

Key takeaways:

  • The project aims to create a low-cost tool for the visually impaired to receive live descriptions of scenes using an ESP32-CAM and AI model.
  • Current limitations include the need for a web page to be open on a cellphone for descriptions and the lack of security for the proof-of-concept.
  • Alternative products are expensive, with prices ranging from $300 to $5900, but they offer varying levels of functionality and accessibility.
  • Future improvements could involve using higher quality cameras and integrating the system into glasses for better usability.
View Full Article

Comments (0)

Be the first to comment!