Waste Classification using Transfer Learning with Convolutional Neural Networks

Abstract With the aim to tackle issue of waste classification for different categories misspend substances, authors, with a limited availability dataset have processed highly accurate model classify garbage into 7 using CompostNet dataset. Experiments were carried out on pre-trained models MobileNetV2, ResNet34 and Densenet121 model, previously trained ImageNet The accuracies obtained 96.42%, 96.27% 96.273% respectively Densenet121, mobilenetv2 resnet34 models. Within 60 epochs, neural network accurately categorizes materials provided in input image. results experiments are compared other previous work done same field. applications conducted this research aims at providing better categorization also follows United Nations goal Responsible Consumption Production towards sustainable development.