Correlation-guided decoding strategy for low-resource Uyghur scene text recognition

Abstract Currently, most state-of-the-art scene text recognition methods are based on the Transformer architecture and rely on pre-trained large language models. However, these pre-trained models are primarily designed for resource-rich languages and exhibit limitations when applied to low-resource...

Täydet tiedot

Bibliografiset tiedot

Päätekijät:	Miaomiao Xu, Jiang Zhang, Lianghui Xu, Wushour Silamu, Yanbing Li
Aineistotyyppi:	Artigo
Kieli:	Inglês
Julkaistu:	Springer 2024-11-01
Sarja:	Complex & Intelligent Systems
Aiheet:	Scene text recognition Low-resource Uyghur Correlation-guided decoding strategy Hybrid encoding
Linkit:	https://doi.org/10.1007/s40747-024-01689-5
Tagit:	Lisää tagi Ei tageja, Lisää ensimmäinen tagi!