K
Kaiming He
Researcher at Facebook
Publications - 140
Citations - 440003
Kaiming He is an academic researcher from Facebook. The author has contributed to research in topics: Object detection & Image segmentation. The author has an hindex of 89, co-authored 135 publications receiving 272091 citations. Previous affiliations of Kaiming He include The Chinese University of Hong Kong & Microsoft.
Papers
More filters
Journal ArticleDOI
Image Completion Approaches Using the Statistics of Similar Patches
Kaiming He,Jian Sun +1 more
TL;DR: This paper observes that if the authors match similar patches in the image and obtain their offsets (relative positions), the statistics of these offsets are sparsely distributed and can be incorporated into both matching-based and graph-based methods for image completion.
Proceedings ArticleDOI
Computing nearest-neighbor fields via Propagation-Assisted KD-Trees
Kaiming He,Jian Sun +1 more
TL;DR: A novel propagation search method for kd-trees where the tree nodes checked by each query are propagated from the nearby queries, which not only avoids the time-consuming backtracking in traditional tree methods, but is more accurate.
Posted Content
Detecting and Recognizing Human-Object Interactions
TL;DR: In this paper, a human-centric approach is proposed to detect triplets in challenging everyday photos by predicting an action-specific density over target object locations based on the appearance of a detected person.
Proceedings ArticleDOI
Masked Autoencoders As Spatiotemporal Learners
TL;DR: It is shown that the MAE method can learn strong representations with almost no inductive bias on spacetime, and spacetime- agnostic random masking performs the best, and the general framework of masked autoencoding can be a unified methodology for representation learning with minimal domain knowledge.
Posted Content
Instance-sensitive Fully Convolutional Networks
TL;DR: This paper develops FCNs that are capable of proposing instance-level segment candidates that do not have any high-dimensional layer related to the mask resolution, but instead exploits image local coherence for estimating instances.