Image-Based Barcode Detection and Recognition to Assist Visually Impaired Persons

Millions of people worldwide suffer from visual impairment or blindness. One of the major limitations caused by the decrease in visual abilities is the difficulties encountered during shopping. In this paper, we propose a barcode detection and recognition method for a shopping assistant. The proposed method is capable of extracting the essential product information (e.g., ingredients, source, calories, etc.) from the detected barcode region. Suggestions of human-system interaction are provided to enhance the user experience. The shopping assistant facilitates the process of grocery shopping and increases the independency of visually impaired persons. To achieve a real-time detection and recognition, we build the detection framework upon a state-of-the-art parallel segment detector that takes advantages of the property that barcode consists of a group of parallel lines. Barcode recognition is accomplished utilizing a robust phone application that provides the detailed information of the product in real-time. As demonstrated by the experimental results, the proposed method is effective, robust, and shows great potential to be expanded into an assistive system for visually impaired persons.