HTTD: A Hierarchical Transformer for Accurate Table Detection in Document Images DOI Creative Commons
Mahmoud SalahEldin Kasem, Mohamed Mahmoud,

Bilel Yagoub

и другие.

Mathematics, Год журнала: 2025, Номер 13(2), С. 266 - 266

Опубликована: Янв. 15, 2025

Table detection in document images is a challenging problem due to diverse layouts, irregular structures, and embedded graphical elements. In this study, we present HTTD (Hierarchical Transformer for Detection), cutting-edge model that combines Swin-L backbone with advanced Transformer-based mechanisms achieve superior performance. addresses three key challenges: handling including historical modern structures; improving computational efficiency training convergence; demonstrating adaptability non-standard tasks like medical imaging receipt detection. Evaluated on benchmark datasets, achieves state-of-the-art results, precision rates of 96.98% ICDAR-2019 cTDaR, 96.43% TNCR, 93.14% TabRecSet. These results validate its effectiveness efficiency, paving the way analysis data digitization tasks.

Язык: Английский

What Lies Ahead for Marketing in Wholesale and Retailing What Lies Ahead for Marketing in Wholesale and Retailing DOI
Theodore Tarnanidis

Advances in marketing, customer relationship management, and e-services book series, Год журнала: 2024, Номер unknown, С. 354 - 363

Опубликована: Авг. 21, 2024

This chapter analyzes the future of marketing science in wholesaling and retailing. Based on findings academic literature, it can argued that retailing will be characterized by a blend data-driven insights, technological innovation, personalized experiences, focus sustainability ethics. Success this evolving landscape require agility, adaptability, deep understanding consumer preferences behaviors.

Язык: Английский

Процитировано

7
HTTD: A Hierarchical Transformer for Accurate Table Detection in Document Images DOI Creative Commons
Mahmoud SalahEldin Kasem, Mohamed Mahmoud,

Bilel Yagoub

и другие.

Mathematics, Год журнала: 2025, Номер 13(2), С. 266 - 266

Опубликована: Янв. 15, 2025

Table detection in document images is a challenging problem due to diverse layouts, irregular structures, and embedded graphical elements. In this study, we present HTTD (Hierarchical Transformer for Detection), cutting-edge model that combines Swin-L backbone with advanced Transformer-based mechanisms achieve superior performance. addresses three key challenges: handling including historical modern structures; improving computational efficiency training convergence; demonstrating adaptability non-standard tasks like medical imaging receipt detection. Evaluated on benchmark datasets, achieves state-of-the-art results, precision rates of 96.98% ICDAR-2019 cTDaR, 96.43% TNCR, 93.14% TabRecSet. These results validate its effectiveness efficiency, paving the way analysis data digitization tasks.

Язык: Английский

Процитировано

0