Implementasi Metode Hybrid MobileNet dan Support Vector Machine untuk Klasifikasi Citra Sampah

This research aims to develop a litter image classification system using the
Hybrid MobileNet method as a feature extraction and Support Vector Machine (SVM)
as a classification algorithm. By utilizing the ability of MobileNet in extracting visual
features from images and SVM in classifying data optimally, this system is expected to
improve the accuracy of garbage classification. The data used comes from a waste
dataset consisting of several categories such as organic, plastic, glass, metal, paper,
plastic and mixed waste. This research uses a dataset with 3200 images divided into
training and test data. The test results showed that the combination of these two
methods resulted in a classification accuracy of 89%, with other performance metrics
such as precision, recall, and F1-score consistently above 85%. The implementation of
this method is expected to make a real contribution to automated waste management
and support recycling efforts.