Recognizing food is an everyday challenge for blind and visually impaired people: Without visual cues, many products are difficult or impossible to identify, which significantly limits their independence in everyday life. The researchers at TTZ Günzburg are providing a remedy: The “Food Container Identifier” project uses an AI model that can recognize food containers with the help of a camera.
AI technology with perspective
The AI model was developed on an Nvidia Jetson Nano; a compact and energy-efficient computer that was specially developed for applications in the field of artificial intelligence and machine learning. Student assistants from the TTZ Günzburg photographed various foods to create a comprehensive database for training the AI. The images were taken under various conditions - different perspectives, backgrounds and lighting conditions - in order to make the model as robust and efficient as possible.
The system's voice output tells users the name of the identified object, which significantly improves the independence of blind people when shopping or organizing their supplies. In this way, the Food Container Identifier offers practical support in everyday life and increases the independence of the people concerned.
Further projects planned for everyday AI applications
At the TTZ Günzburg, further applications are to be developed beyond the current use case to demonstrate the potential of modern AI technologies in practical applications. For the researchers at the TTZ, projects of this kind are not just technological achievements: “Every project brings with it new insights and opportunities aimed at using technology in a meaningful way to solve specific problems and have a positive impact on society,” says Prof. Dr. Philipp Brune.
The TTZ Günzburg
With a focus on artificial intelligence (AI) and data analytics, the TTZ aims to sustainably promote and strengthen the regional economy on behalf of the Bavarian state government. The TTZ actively supports companies and their employees by implementing innovation projects. Companies can contact the TTZ experts on the website (www.hnu.de/ttzguenzburg) or by email at info-ttz@hnu.de.
