Environment friendly Plastic Garbage Classification Using Convolution Neural Network

Plastic pollution poses a global threat, necessitating innovative waste management solutions. This research leverages Convolutional Neural Networks (CNNs) to precisely classify plastic waste, with a focus on comparing VGG16 and RESNET50 models. The primary aim is to develop robust models contributing to sustainable waste management practices.Curated datasets for supervised learning are employed, utilizing CNN architectures VGG16 and RESNET50 for feature extraction. The study explores integrating graph and recurrent neural networks for enhanced discernment of spatial and temporal plastic features, addressing the complex nature of plastic waste. The integration of these advanced models with theoretical frameworks improves waste sorting accuracy. Experimental results showcase the effectiveness of CNN in plastic waste classification. VGG16 achieves a commendable accuracy of 75%, while RESNET50 outperforms with 81%. The increase in accuracy highlights RESNET50’s superior feature learning capabilities. The choice between these models significantly impacts automated plastic waste sorting systems.