Few Could Be Better Than All: Feature Sampling and Grouping for Scene Text Detection.
Jingqun TangWenqing ZhangHongye LiuMingkun YangBo JiangGuanglong HuXiang BaiPublished in: CVPR (2022)
Keyphrases
- text detection
- urban scenes
- outdoor scenes
- natural scene images
- object recognition
- d scene
- scene understanding
- video analysis
- text information
- connected components
- three dimensional
- video sequences
- urban environments
- image features
- moving objects
- range data
- structure from motion
- input image
- image sequences
- aerial images
- vanishing points
- planar surfaces