Search Results for author:
Found 3 papers, 1 papers with code
MVP-Human Dataset for 3D Human Avatar Reconstruction from Unconstrained Frames
In this paper, we consider a novel problem of reconstructing a 3D human avatar from multiple unconstrained frames, independent of assumptions on camera calibration, capture space, and constrained actions.
Augmented Bi-path Network for Few-shot Learning
Finally, the model learns to compare global and local features separately, i. e., in two paths, before merging the similarities.
Quadruplet Network with One-Shot Learning for Fast Visual Object Tracking
In this paper, we propose a new quadruplet deep network to examine the potential connections among the training instances, aiming to achieve a more powerful representation.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
