K
Kai Han
Researcher at University of Bristol
Publications - 55
Citations - 1437
Kai Han is an academic researcher from University of Bristol. The author has contributed to research in topics: Cluster analysis & Photometric stereo. The author has an hindex of 15, co-authored 55 publications receiving 645 citations. Previous affiliations of Kai Han include University of Hong Kong & University of Oxford.
Papers
More filters
Proceedings ArticleDOI
Learning to Discover Novel Visual Categories via Deep Transfer Clustering
TL;DR: In this article, the authors extend Deep Embedded Clustering to a transfer learning setting and propose a method to estimate the number of classes in the unlabeled data, using knowledge from the known classes.
Proceedings ArticleDOI
Contrastive Learning based Hybrid Networks for Long-Tailed Image Classification
TL;DR: In this article, the authors propose a hybrid network structure composed of a supervised contrastive loss to learn image representations and a cross-entropy loss for learning classifiers, where the learning is progressively transited from feature learning to the classifier learning to embody the idea that better features make better classifiers.
Proceedings ArticleDOI
SCNet: Learning Semantic Correspondence
Kai Han,Rafael Sampaio de Rezende,Bumsub Ham,Kwan-Yee K. Wong,Minsu Cho,Cordelia Schmid,Jean Ponce +6 more
TL;DR: SCNet as discussed by the authors uses region proposals as matching primitives, and explicitly incorporates geometric consistency in its loss function to establish semantic correspondences between images depicting different instances of the same object or scene category.
Book ChapterDOI
PS-FCN: A Flexible Learning Framework for Photometric Stereo
TL;DR: Deep fully convolutional network that takes an arbitrary number of images of a static object captured under different light directions with a fixed camera as input, and predicts a normal map of the object in a fast feed-forward pass, and can handle multiple images and light directions in an order-agnostic manner is proposed.
Proceedings ArticleDOI
Self-Calibrating Deep Photometric Stereo Networks
TL;DR: An uncalibrated photometric stereo method for non-Lambertian scenes based on deep learning that can effectively take advantage of intermediate supervision, resulting in reduced learning difficulty compared to a single-stage model is proposed.