DeepWSD: Projecting Degradations in Perceptual Space to Wasserstein Distance in Deep Feature Space.
Xingran LiaoBaoliang ChenHanwei ZhuShiqi WangMingliang ZhouSam KwongPublished in: ACM Multimedia (2022)
Keyphrases
- feature space
- lower dimensional
- low dimensional
- input space
- high dimensional
- dimensionality reduction
- higher dimensional
- principal component analysis
- mean shift
- dissimilarity measure
- distance measure
- image quality
- euclidean distance
- distance metric
- data points
- classification accuracy
- kernel space
- high dimensional feature space
- original data
- lower dimension
- space time
- distance function
- high dimensional data
- kernel function
- search space
- learning algorithm
- vector space
- high dimensionality
- parameter space
- human perception
- shortest path
- geodesic distance
- training samples
- support vector machine
- linearly separable
- training set
- similarity measure