Abstract
We address the problem of weakly supervised semantic segmentation. The training images are labeled only by the classes they contain, not by their location in the image. On test images instead, the method must predict a class label for every pixel. Our goal is to enable segmentation algorithms to use multiple visual cues in this weakly supervised setting, analogous to what is achieved by fully supervised methods. However, it is difficult to assess the relative usefulness of different visual cues from weakly supervised training data. We define a parametric family of structured models, were each model weights visual cues in a different way. We propose a Maximum Expected Agreement model selection principle that evaluates the quality of a model from the family without looking at superpixel labels. Searching for the best model is a hard optimization problem, which has no analytic gradient and multiple local optima. We cast it as a Bayesian optimization problem and propose an algorithm based on Gaussian processes to efficiently solve it. Our second contribution is an Extremely Randomized Hashing Forest that represents diverse superpixel features as a sparse binary vector. It enables using appearance models of visual classes that are fast at training and testing and yet accurate. Experiments on the SIFT-flow dataset show a significant improvement over previous weakly supervised methods and even over some fully supervised methods.
Original language | English |
---|---|
Title of host publication | Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on |
Publisher | Institute of Electrical and Electronics Engineers |
Pages | 845-852 |
Number of pages | 8 |
ISBN (Electronic) | 978-1-4673-1227-1 |
ISBN (Print) | 978-1-4673-1226-4 |
DOIs | |
Publication status | Published - 1 Jun 2012 |
Keywords / Materials (for Non-textual outputs)
- Bayes methods
- Gaussian processes
- image segmentation
- learning (artificial intelligence)
- optimisation
- Bayesian optimization problem
- Gaussian process
- SIFT-flow dataset
- appearance model
- diverse superpixel feature
- extremely randomized hashing forest
- hard optimization problem
- maximum expected agreement model selection
- multiple visual cues
- semantic segmentation
- sparse binary vector
- weakly supervised structured output learning
- Image segmentation
- Kernel
- Measurement
- Optimization
- Semantics
- Training
- Visualization