https://libeldoc.bsuir.by/handle/123456789/60622
Title: | Methods of low data image classification with neural network |
Authors: | Chen, Y. M. |
Keywords: | материалы конференций;Image classification;Deep learning;Zero-shot learning;Few-shot learning;CLIP;Convolutional Neural Networks |
Issue Date: | 2025 |
Publisher: | БГУИР |
Citation: | Chen, Y. M. Methods of low data image classification with neural network / Y. M. Chen // Технологии передачи и обработки информации : материалы Международного научно-технического семинара, Минск, апрель 2025 г. / Белорусский государственный университет информатики и радиоэлектроники ; редкол.: В. Ю. Цветков [и др.]. – Минск, 2025. – С. 162–164. |
Abstract: | Image classification has evolved from relying on handcrafted feature engineering to a data-driven deep learning paradigm, with recent breakthroughs extending its capabilities to zeroshot and few-shot learning scenarios. This paper introduces emerging paradigms for low-data scenarios: zero-shot learning methods, including multimodal models (e.g., CLIP) that align visual and textual embeddings for open vocabulary classification and generative frameworks for synthesizing features of unseen classes; and (2) few-shot learning strategies, such as meta-learning (e.g., MAML), metric-based networks (e.g., ProtoNet), and fast adaptation techniques (e.g., Tip-Adapter) that leverage pre-trained knowledge for fast adaptation. While these methods reduce the reliance on labeled data, challenges remain in domain adaptation, fine-grained classification, and computational efficiency. |
URI: | https://libeldoc.bsuir.by/handle/123456789/60622 |
Appears in Collections: | Технологии передачи и обработки информации : материалы Международного научно-технического семинара (2025) |
File | Description | Size | Format | |
---|---|---|---|---|
Chen_Methods.pdf | 411.23 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.