Optimizing Furniture Assembly: A CNN-based Mobile Application for Guided Assembly and Verification

Abstract

Building products based o written instructions can be challenging for individuals, often leading to difficulties during the assembly process. This project presents a mobile application that guides users through the assembly process and verifies assembly correctness using image comparison techniques. The application features manufacturer and consumer modes, allowing manufacturers to add assembly steps and users to access provided assembly instructions for a given product. Cross-platform compatibility is achieved with React-Native and Expo Go, while image matching is implemented by incorporating computer vision and machine learning techniques, specifically utilizing a Siamese Network architecture. Amazon Web Services (AWS) supports content delivery and storage. An IKEA table and kids chair assembly serves as a demonstrative example. By offering step-by-step guidance and feedback based on similarity indexes, the application optimizes the efficiency and accuracy of furniture assembly. The project showcases the system’s practicality, particularly in enhancing the assembly process for various furniture items, exemplified by the IKEA products. Keywords: Computer vision, Web services, Object detection, Network architecture, User experience, Robustness, Mobile applications