Medical Waste Classification Using Convolutional Neural Network

Abstract

Medical waste disposal is a significant issue in developing countries like Thailand. It poses a persistent public health challenge as it leads to contamination of the environment and the spread of infectious diseases. This study aims to address this challenge by using a deep learning model to categorize different types of medical waste, including vials, masks, syringes, gloves, cotton, bandages, and IV tubes. Transfer learning method was employed to enhance the classification process. The study utilized the EfficientNet_b7 model and evaluated its performance based on accuracy, precision, recall, and F1 score. The results showed that with transfer learning, EfficientNet_b7 achieved a classification accuracy of 99% for both the training and testing datasets. Although there was a decline in accuracy, particularly for the syringe class, pretrained CNNs significantly improved the efficiency and accuracy of medical waste classification. Consequently, this proposed CNN model can serve as a viable alternative to conventional methods for classifying medical waste. By implementing these approaches, the efficiency of waste classification is improving, leading to a reduction in the costs associated with manual classification. This promotes sustainable waste management practices, which in turn contribute to the overall health of ecosystems and human well-being.