ScrapNet: An Efficient Approach to Trash Classification

As people have become more aware of their actions and how they affect surroundings, realized the dire state environment. a result, recycling movement has gained momentum as measure to save it. Contemporarily, industry not seen major shift problems that existed decades earlier persist. Trash classification is at core these problems, because if you can’t classify it, obviously recycle Manual often leads misclassification humans or judge things based on experiences, knowledge, just absolute facts. Additionally, waste be sorted toxic, being in direct contact may physically harmful involved. Until solution found for this problem, won’t par with rise culture. Thus, problem we set out solve; paper proposes Deep Learning model EfficientNet Architecture can different kinds trash an accuracy high enough make it viable while using comparatively lower number parameters than existing methods. We achieved 98% TrashNet dataset, standard data classification, outperformed all models. no large dataset varied images was present, created new 8135 by combining various datasets standardized them, achieving 92.87% B3.