Title
PyrBoxes: An Efficient Multi-scale Scene Text Detector with Feature Pyramids
Abstract
•An efficient and effective model is proposed for multi-scale scene text detection.•A grouped pyramid module is used to group basic layers for more accurate detection.•The pyramid module includes a top-down partition and a bottom-up group.•PyrBoxes achieves the state-of-the-art performance with a significantly faster speed.•The generality of PyrBoxes is verified by a newly constructed Chinese text dataset.
Year
DOI
Venue
2019
10.1016/j.patrec.2019.04.022
Pattern Recognition Letters
Keywords
Field
DocType
41A05,41A10,65D05,65D17
Pattern recognition,Inference,Append,Artificial intelligence,Pyramid,Partition (number theory),Detector,Generality,Semantics,Recursion,Mathematics
Journal
Volume
ISSN
Citations 
125
0167-8655
1
PageRank 
References 
Authors
0.43
0
4
Name
Order
Citations
PageRank
Fenfen Sheng110.43
Zhineng Chen219225.29
Wei Zhang341.98
Bo Xu411127.31