Using line segments to train multi-stream stacked autoencoders for image classification.
Xue-Song TangKuangrong HaoHui WeiYongsheng DingPublished in: Pattern Recognit. Lett. (2017)
Keyphrases
- line segments
- multi stream
- image classification
- audio visual speech recognition
- hough transform
- hidden markov models
- audio visual
- bag of words
- denoising
- digital pictures
- visual features
- straight line
- feature extraction
- image representation
- ellipse detection
- circular arcs
- straight line segments
- aerial images
- visual information
- image features
- intersection points
- feature selection
- perspective images
- multiscale
- high dimensional
- low level