Recognizing Aluminum Beverage Cans from Waste Mixtures Based on Densenet121-CNN Model: Deep Learning Methodology

Waste mixture classification and recycling a major contributor to the recycling economy and sustainable environment. Recognizing valuable and expensive elements from waste mixtures images for recycling purposes is a major area of interest within the field of artificial intelligence. One of the interesting and important waste recycling issues is aluminum production from aluminum beverage cans. Therefore, recognizing aluminum beverage cans from solid and bottle waste mixtures is an interesting and important research problem. In this paper, a proposed deep learning model called DenseNet121-CNN is designed and developed to classify solid and bottle wastes to recognize aluminum beverage cans. The proposed model has been tested and trained using a benchmark dataset that consisting of 4825 bottle waste images. The experimental and validation results proved the efficiency of DenseNet 121-CNN in recognizing aluminum beverage can images, where, it achieved training and validation accuracy reached 99.9% and 99.7% respectively. The proposed model results also confirmed the superiority of the DenseNet 121-CNN compared to other proposed architectures such as Xception-CNN and ResNet50-CNN as well as other state-of-the-art machine and deep learning methodologies.