PENERAPAN ARSITEKTUR JST DALAM DEEP LEARNING UNTUK MENINGKATKAN AKURASI KLASIFIKASI GAMBAR DENGAN AUTOENCODER

Bahasa Indonesia

Abstract

Abstract - The development of artificial intelligence technology, particularly deep learning, has made significant contributions to digital image processing across various fields such as medicine, security, and manufacturing industries. This study aims to implement the autoencoder method within an Artificial Neural Network (ANN) architecture to optimally enhance image classification accuracy. The autoencoder is employed as an unsupervised learning technique to extract essential and relevant features from input images before passing them to the classification layer. The training process was carried out using a carefully curated image dataset, and the model was evaluated to measure classification performance based on accuracy, precision, and recall. The experimental results show that integrating an autoencoder into the ANN architecture can improve feature extraction efficiency, reduce noise, and deliver more accurate and consistent classification results compared to conventional approaches. This research demonstrates that the autoencoder can serve as a vital component in modern deep learning-based classification systems.