Blind Bite: Food Ingredients Identification and Assistance App for Visually Impaired

Visually impaired individuals face significant challenges in identifying food ingredients, impacting their independence and quality of life. Traditional methods, such as reading labels or visually recognizing ingredients, are often not feasible for these users. Existing technological solutions, like barcode scanning, are limited to packaged items and fail to address fresh produce or unpackaged ingredients, highlighting a gap in current offerings. Current models also struggle with inconsistency in recognition accuracy under varying conditions, high computational requirements, and lack of customization options for userspecific needs. To address these gaps, we propose a smartphone-based application utilizing advanced image recognition techniques, such as Vision Transformers and Convolutional Neural Networks (CNNs), optimized for real-time performance on mobile devices. The application adapts to various lighting conditions, enhancing reliability and usability in diverse settings. Designed with accessibility in mind, the app incorporates voiceactivated controls and text-to-speech functionality to provide a seamless, hands-free user experience for visually impaired individuals. The proposed solution aims to provide a comprehensive and accessible tool for food ingredient recognition, enhancing the autonomy of visually impaired users and enabling informed food choices in their daily lives.