DeepWSD: Projecting Degradations in Perceptual Space to Wasserstein Distance in Deep Feature Space.
Xingran LiaoBaoliang ChenHanwei ZhuShiqi WangMingliang ZhouSam KwongPublished in: CoRR (2022)
Keyphrases
- feature space
- lower dimensional
- low dimensional
- input space
- high dimensional feature space
- higher dimensional
- data points
- dimensionality reduction
- linearly separable
- high dimensional
- mean shift
- space time
- distance measure
- dissimilarity measure
- principal component analysis
- classification accuracy
- search space
- original data
- kernel space
- image processing
- machine learning
- feature vectors
- feature set
- euclidean distance
- low level
- support vector machine
- pairwise distances
- lower dimension
- training set
- riemannian manifolds
- relative position
- human perception
- euclidean space
- image representation
- parameter space
- high dimensional data
- distance metric