Kriswono, Riski (2024) Implementasi Convolutional Neural Network (cnn) Dengan Arsitektur Densenet Untuk Klasifikasi Sampah Anorganik. Other thesis, Universitas Islam Riau.

Text 183510542.pdf - Submitted Version Restricted to Registered users only Download (2MB) | Request a copy

Abstract

The management of inorganic waste is one of the main challenges in the modern era considering its nature that is difficult to decompose naturally and its impact on the environment. This research aims to implement the Convolutional Neural Network (CNN) method with DenseNet-121, DenseNet-169, and DenseNet-201 architectures in the classification of inorganic waste types. The dataset used consists of five classes, namely cardboard, glass, metal, paper, and plastic, with a total of 3,537 images obtained from the Kaggle platform. The research process includes data preprocessing, data augmentation, model training with hyperparameter configuration, and evaluation using confusion matrix to measure accuracy, precision, recall, and f1-score. The results showed that DenseNet-201 gave the best performance with 90% accuracy, followed by DenseNet-169 (89%) and DenseNet-121 (88%). The web-based classification system developed using the Flask framework allows users to upload litter images and receive classification results automatically

Actions (login required)

View Item