Split-net: Dual transformer encoder with splitting scene text image for script identification

doi:10.1016/j.patrec.2025.05.026

Split-net: Dual transformer encoder with splitting scene text image for script identification

2025-06-04

0

PRE

AI

A

Ayush Roy

P

Palaiahnakote Shivakumara

DOI：10.1016/j.patrec.2025.05.026

Original

Original request for help

Save

Abstract

En 中文

• Proposing a dual encoder for generating local features. • A novel Attention Module to integrate the transformer blocks of M-ViT encoders. • Gradient-based feature fusion to integrate the extracted features. • Experimental results show that the proposed method is better than the state-of-the-art.

Split-net: Dual transformer encoder with splitting scene text image for script identification

Abstract

Journal

Organization