Puteri Nurul Ma'rifah, Moechammad Sarosa, Erfan Rohadi
Inefficiency in waste management contributes to the rising amount of pollution in the community, causing the public demand for more proper waste management and classification. Sorting or classifying waste is the beginning of the waste recycling process, which can help reduce the amount of garbage in the environment. However, compounded with the lack of waste sorting awareness due to minimum public education on waste management, the waste sorting system is still conducted manually using human power. Thus, it is necessary to have a waste sorting or classification system to encourage people to manage their waste properly. This study aims to design a tool to detect the waste types and classify them into three categories; metal, paper, and plastic waste. The system can recognize the shape of the waste image using deep learning methods developed using the Faster R-CNN with Resnet50 as the network architecture. This research began with collecting 250 metal, paper, and plastic waste datasets used for training data and test data in the testing process. The test consisted of 6 trials. In each trial, the training process ran up to 3,000, 4,000, 5,000, and 8,000 steps, where each had a loss function parameter. The test results show that increasing the number of steps during the exercise increases the F1 score, but the maximum number of steps with an increase in the F1 score is 5,000 by 91%. © 2023 IEEE.
State Polytechnic of Malang, Department of Electrical Engineering, Malang, Indonesia; State Polytechnic of Malang, Department of Information Technology, Malang, Indonesia