J
Jian Sun
Researcher at Xi'an Jiaotong University
Publications - 394
Citations - 356427
Jian Sun is an academic researcher from Xi'an Jiaotong University. The author has contributed to research in topics: Computer science & Object detection. The author has an hindex of 109, co-authored 360 publications receiving 239387 citations. Previous affiliations of Jian Sun include French Institute for Research in Computer Science and Automation & Tsinghua University.
Papers
More filters
Journal Article
Normal and feature approximations from noisy point clouds
Tamal K. Dey,Jian Sun +1 more
TL;DR: In this paper, the authors consider the problem of approximating normal and feature sizes of a surface from point cloud data that may be noisy and provide new algorithms for practical and reliable norm and feature approximations.
Journal ArticleDOI
Approximating cycles in a shortest basis of the first homology group from point data
Tamal K. Dey,Jian Sun,Yusu Wang +2 more
TL;DR: An algorithm to compute a set of cycles from a point data that presumably sample a smooth manifold M ⊂ ℝ d to approximate a shortest basis of the first homology group H 1 (M) over coefficients in the finite field ℤ 2 .
Patent
Enhanced user eye gaze estimation
TL;DR: In this paper, a plurality of images of a user's eye are acquired and an enhanced user eye gaze is estimated by narrowing a database of eye information and corresponding known gaze lines to a subset of the eye information having gaze lines corresponding to a gaze target area.
Proceedings ArticleDOI
Gromov-Hausdorff Approximation of Filament Structure Using Reeb-type Graph
Frédéric Chazal,Jian Sun +1 more
TL;DR: It is proved that filamentary structures that can be seen as topological metric graphs can be approximated, with respect to the Gromov-Hausdorff distance by well-chosen Reeb graphs (and some of their variants) and an efficient and easy to implement algorithm to compute such approximations is provided.
Posted Content
WeightNet: Revisiting the Design Space of Weight Networks
TL;DR: The WeightNet, composed entirely of (grouped) fully-connected layers, is used to directly output the convolutional weight, and outperforms existing approaches on both ImageNet and COCO detection tasks, achieving better Accuracy-FLOPs and Accuracy-Parameter trade-offs.