Accelerating Transformer-Based Scene Text Detection and Recognition via Token Pruning.
Sergi Garcia-BordilsDimosthenis KaratzasMarçal RusiñolPublished in: ICDAR (6) (2023)
Keyphrases
- text detection
- scene text
- urban scenes
- object recognition
- natural scene images
- outdoor scenes
- scene understanding
- video analysis
- text information
- connected components
- video sequences
- scene images
- d scene
- image sequences
- urban environments
- structure from motion
- single image
- three dimensional
- complex background
- range data
- input image
- complex scenes
- video data
- natural images
- image classification
- moving objects