Waste Classification for Sustainable Development Using Image Recognition with Deep Learning Neural Network Models

The proper handling of waste is one the biggest challenges modern society. Municipal Solid Waste (MSW) requires categorization into a number types, including bio, plastic, glass, metal, paper, etc. most efficient techniques proposed by researchers so far include neural networks. In this detailed summarization was made existing deep learning that have been to classify waste. This paper proposes an architecture for classification litter categories specified in benchmark approaches. used EfficientNet-B0. These are compound-scaling based models Google pretrained on ImageNet and accuracy 74% 84% top-1 over ImageNet. research EfficientNet-B0 model tuning images specific particular demographic regions classification. type transfer provides customized classification, highly optimized region. It shown such had comparable EfficientNet-B3, however, with significantly smaller parameters required B3 model. Thus, technique achieved efficiency order 4X terms FLOPS. Moreover, it resulted improvised classifications as result fine-tuning region-wise images.