Platform to Help CNN Training in Manufacturing Teaching Environments

This research-to-practice full paper presents a platform developed to support AI teaching, training, and evaluation. Artificial intelligence (AI) has revolutionized several fields, including pattern recognition, natural language processing, and computer vision. In this context, convolutional neural networks (CNNs) have been effectively utilized to detect objects in images. In industry, AI has the potential to optimize processes and increase efficiency, becoming an essential factor in the evolution of manufacturing systems. However, the training of AI models is complex and requires a sophisticated level of knowledge for data manipulation and an understanding of model training. To simplify this process, a web-based platform has been proposed to facilitate the training of YOLO networks. The platform was developed via the Streamlit library, which provides an interface to guide users through model training. The platform aims to lead users through six steps to obtain and deploy a model directly through the platform, enabling performance evaluation and metric tracking.