Inspiration Many small grocery shopkeepers lack barcode-based inventory systems due to cost and implementation barriers. We wanted to create an AI-driven solution that simplifies billing without requiring barcodes, making the process more efficient and accessible.
What it does Scannerless Biller uses image recognition and OCR technology to identify products from images and generate bills automatically. Shopkeepers can simply take a picture of the items, and the system extracts product names, retrieves prices from a database, and creates a bill in real-time.
How we built it Backend: Built using Flask for efficient processing and API handling. Machine Learning: Utilized MobileNetV2 for image classification and EasyOCR for text extraction. Database: Used MongoDB to store product details and manage billing records. Frontend: Designed a simple and intuitive interface for shopkeepers to upload images and view generated bills. Challenges we ran into Achieving high accuracy in recognizing different product images and text under varying lighting conditions. Optimizing OCR results to correctly interpret product names from labels. Managing real-time data retrieval efficiently for a smooth billing experience. Accomplishments that we're proud of Successfully built a functional prototype that eliminates the need for barcode scanners. Integrated MobileNetV2 and OCR seamlessly for product recognition. Developed a solution that can be easily adopted by small businesses. What we learned The importance of dataset quality in improving image recognition accuracy. Techniques for fine-tuning MobileNetV2 and optimizing OCR outputs. Managing database queries efficiently to enhance system performance. What's next for Scannerless Biller Expanding the product database to improve recognition capabilities. Enhancing the UI/UX for a smoother user experience. Exploring cloud-based deployment for scalability. Adding multilingual OCR support for broader adoption.
Log in or sign up for Devpost to join the conversation.