GISCA: Gradient-Inductive Segmentation Network With Contextual Attention for Scene Text Detection.
Meng CaoYuexian ZouDongming YangChao LiuPublished in: IEEE Access (2019)
Keyphrases
- text detection
- outdoor scenes
- urban scenes
- natural scene images
- text information
- video analysis
- edge detection
- level set
- contextual information
- connected components
- object recognition
- video sequences
- segmentation algorithm
- image regions
- image segmentation
- multiple objects
- single image
- video database
- multiscale
- point cloud
- complex background
- photorealistic
- input image