Diffusion-based inpainting of incomplete Euclidean distance matrices of trajectories generated by a fractional Brownian motion.
Alexander LobashevKirill PolovnikovPublished in: CoRR (2024)
Keyphrases
- euclidean distance
- fractional brownian motion
- long range
- frobenius norm
- distance metric
- non stationary
- distance measure
- dimensionality reduction
- distance function
- data points
- similarity measure
- mahalanobis distance
- fractal dimension
- feature vectors
- image restoration
- manhattan distance
- diffusion process
- anisotropic diffusion
- geodesic distance
- singular value decomposition
- denoising
- euclidean norm
- financial markets
- image processing
- long range dependence
- random fields
- edge detection
- co occurrence
- nearest neighbor
- pairwise
- pattern recognition
- feature extraction