CNN-Based Food Image Segmentation Without Pixel-Wise Annotation.
Wataru ShimodaKeiji YanaiPublished in: ICIAP Workshops (2015)
Keyphrases
- pixel wise
- image segmentation
- spatial information
- deformable models
- markov random field
- foreground objects
- active contours
- multiscale
- pixel level
- background subtraction
- object segmentation
- appearance model
- image processing
- template matching
- graph cuts
- level set
- segmentation algorithm
- active learning
- image retrieval
- higher order
- expectation maximization
- multi view
- image analysis
- graph structure
- energy function
- three dimensional
- maximum likelihood