Improving Unsupervised Domain Adaptive Re-Identification Via Source-Guided Selection of Pseudo-Labeling Hyperparameters.
Fabian DubourvieuxAngélique LoeschRomaric AudigierSamia AinouzStéphane CanuPublished in: IEEE Access (2021)
Keyphrases
- hyperparameters
- model selection
- cross validation
- closed form
- bayesian inference
- bayesian framework
- em algorithm
- maximum likelihood
- gaussian processes
- noise level
- prior information
- support vector
- random sampling
- gaussian process
- incremental learning
- active learning
- image segmentation
- image processing
- learning algorithm
- machine learning
- data mining