Implementasi Model Convolutional Neural Network dalam Aplikasi Android untuk Identifikasi Limbah Infeksius

Abstract

After the COVID-19 pandemic passed, Indonesian citizens were still strict about using masks because active cases were still found. However, not all Indonesian people are aware that masks are an infectious waste, so after use, they are still disposed of carelessly. Apart from masks, other infectious waste in the form of battery waste which contains hazardous chemicals and food waste potentially to spread infectious diseases, is also dangerous for humans. These kinds of waste are major contributors to global pollution. Research on waste classification has been carried out a lot, but especially for infectious waste has not received much attention from researchers. For this reason, this research is useful to help the public distinguish infectious waste such as used food scraps, masks, and batteries so that they are more careful in disposing of waste. The research started with collecting datasets, which came from combining several infectious waste datasets available on the internet. This is done because there is no publicly available dataset that specifically contains infectious waste. Then, a classification model is created with Convolutional Neural Network (CNN) algorithm which has an accuracy of more than 90%. This algorithm has been widely used in previous studies but has never been used as a model applied to Android applications to classify infectious waste. In this study, the CNN model is applied to Android applications. From this research, an Android application with the CNN algorithm will be produced which can help Indonesians identify infectious waste with an accuracy of 94%.