Image-to-Character-to-Word Transformers for Accurate Scene Text Recognition.
Chuhui XueJiaxing HuangWenqing ZhangShijian LuChanghu WangSong BaiPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2023)
Keyphrases
- multiscale
- input image
- image analysis
- image data
- single image
- image segmentation
- template matching
- image content
- image representation
- image collections
- edge detection
- handwritten words
- pixel values
- region of interest
- image classification
- image retrieval
- scene images
- image features
- image pixels
- computer vision
- high resolution
- similarity measure
- image regions
- image set
- printed text
- high fidelity
- scene text recognition
- test images
- segmentation method
- segmentation algorithm
- visual features
- feature points
- low level