UNSUPERVISED NEURAL NETWORKS FOR LEARNING DISCRIMINATIVE IMAGE REPRESENTATIONS
Mamataliyev DiyorbekNamangan State University, Uzbekistan
2026
ABI
Annotatsiya
Unsupervised neural networks provide an effective approach to learning informative image representations without requiring manually labeled datasets. This thesis examines convolutional neural networks, clustering-based learning, and self-supervised methods for extracting discriminative visual features. These techniques can identify meaningful image patterns and support image clustering, retrieval, classification, and anomaly detection. The study outlines a general framework for unsupervised feature extraction and discusses its potential applications and limitations.
Hali tarjima qilinmagan
Identifikatorlar
Iqtiboslar va manbalar
0 ta iqtibos4 ta foydalanilgan manba