SuSana Distancia is all you need: Enforcing class separability in metric learning via two novel distance-based loss functions for few-shot image classification.
Mauricio Mendez-RuizJorge Gonzalez-ZapataIván Reyes-AmezcuaDaniel Flores-AraizaFrancisco Lopez-TiroAndres Mendez-VazquezGilberto Ochoa-RuizPublished in: CoRR (2023)
Keyphrases
- loss function
- image classification
- metric learning
- pairwise
- distance metric learning
- visual features
- dimensionality reduction
- distance metric
- feature extraction
- image features
- learning tasks
- semi supervised
- euclidean distance
- distance measure
- image representation
- distance function
- visual words
- discriminant analysis
- intra class
- support vector
- graph embedding
- linear discriminant analysis
- multi task
- feature space
- pairwise constraints
- multi label
- sparse representation
- probabilistic model
- learning problems
- high dimensional
- pattern recognition
- multi class
- spectral clustering
- sparse coding
- data mining
- neural network