An Advanced Edge Intelligent Approach for Capsule Defect Recognition Based on CNN Using an Embedded Platform

The capsule, as a carrier of medicine, plays a crucial role in protecting drugs, facilitating gastrointestinal absorption, and enhancing their efficacy. To ensure the quality of capsules, it is essential to identify and remove any defects present on their surface. Traditional inspection methods have often been time-consuming and resource-intensive. However, by leveraging optimized neural network architecture and advanced end-equipment, it is now possible to significantly improve inspection efficiency and resource compatibility. In this study, we propose a novel approach to defect detection on capsules using a lightweight Convolutional Neural Network (CNN) model integrated into an embedded system. This model is designed to adapt to different workload conditions and drug types, thereby enhancing detection efficiency. By exploiting the capabilities of the embedded system, we shift the image processing task to the edge, effectively tackling the inherent problems associated with processing delay, privacy, and load that are prevalent in traditional modes.