Applying Deep Reinforcement Learning Algorithm with ResNet18 CNN for Waste Image Classification

Proper waste classification is essential for effective recycling and reducing environmental harm. This paper presents a new approach to waste classification using Deep Reinforcement Learning (DRL). The model combines convolutional neural networks (CNN) to analyze waste images and a learning system that improves decisions through feedback. By training on a variety of waste image datasets, the model achieves 98.87% accuracy in classifying like electronic waste, organic waste, recycle waste, and trash. Compared to other methods, this approach adapts better to new data and provides more reliable results. This study shows how integration of DRL and ResNet18 CNN can improve waste management systems and support smarter, more efficient recycling.