PositionOCR: Augmenting Positional Awareness in Multi-Modal Models via Hybrid Specialist Integration

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Duan, Chen, Guo, Zhentao, Fu, Pei, Wang, Zining, Zhou, Kai, Yan, Pengfei
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2602.19188