Comprehensive benchmark datasets for Amharic scene text detection and recognition.
Wondimu DikubabDingkang LiangMinghui LiaoXiang BaiPublished in: Sci. China Inf. Sci. (2022)
Keyphrases
- benchmark datasets
- text detection
- scene text
- urban scenes
- natural scene images
- object recognition
- outdoor scenes
- scene understanding
- video analysis
- scene images
- connected components
- text information
- video sequences
- d scene
- three dimensional
- complex background
- single image
- aerial images
- video database
- urban environments
- text lines
- detection method
- multiple objects
- dynamic scenes
- structure from motion
- visual information
- multiple views
- photorealistic
- binary images
- image sequences