A localization/verification scheme for finding text in images and video frames based on contrast independent features and machine learning methods.
Datong ChenJean-Marc OdobezJean-Philippe ThiranPublished in: Signal Process. Image Commun. (2004)
Keyphrases
- machine learning methods
- video frames
- text regions
- video images
- image features
- machine learning approaches
- static images
- input image
- machine learning
- text detection
- video sequences
- video segments
- machine learning algorithms
- scene images
- video dataset
- connected components
- test images
- input video
- video clips
- natural scenes
- video data
- single frame
- successive frames
- object recognition
- document images
- video content
- text information
- optical character recognition
- three dimensional
- image classification
- video footage
- camera movement
- scene text
- complex background
- web images
- feature set
- news video
- key frames
- computer graphics
- natural images
- bounding box
- feature points
- feature vectors
- face recognition
- learning algorithm