Asosiy kontentga oʻtish
proceeding

UNSUPERVISED NEURAL NETWORKS FOR LEARNING DISCRIMINATIVE IMAGE REPRESENTATIONS

Mamataliyev DiyorbekNamangan State University, Uzbekistan
2026
Ochiq kirishGibrid

Litsenziya: cc-by

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