EdgeFlow: Edge-Map Augmented VLM-Based Flowchart Processing for Industrial Requirements Engineering

EdgeFlow enhances Vision Language Models (VLMs) to accurately process flowcharts used in industrial requirements engineering by incorporating edge-map information. This approach addresses the limitations of VLMs in capturing topology-critical visual details, which are crucial for converting static flowchart images into machine-readable models. By augmenting VLMs with edge-map data, EdgeFlow improves the accuracy of flowchart conversion, enabling more effective requirements engineering activities. The proposed method has significant implications for industries that rely heavily on flowcharts, such as manufacturing and software development. According to the researchers, EdgeFlow's ability to accurately capture visual details is a key factor in its success¹. This advancement in flowchart processing has important consequences for the development of more sophisticated industrial automation systems, and highlights the need for continued research into the applications of AI in industrial settings. The ability to accurately convert flowcharts into machine-readable models matters to practitioners because it enables more efficient and automated requirements engineering processes.