Yingda Yin
尹英达
Ph.D. student
School of Computer Science
Peking University
Email: yingda.yin at pku.edu.cn
yingda.yin at gmail.com
|
|
Publications
|
*: equal contribution; † corresponding author(s)
|
|
SAI3D: Segment Any Instance in 3D Scenes
Yingda Yin*,
Yuzheng Liu*,
Yang Xiao*,
Daniel Cohen-Or,
Jingwei Huang,
Baoquan Chen
CVPR 2024
arXiv /
project page /
code
We introduce a zero-shot 3D instance segmentation approach that synergistically leverages geometric priors and semantic cues derived from Segment Anything Model (SAM).
|
|
Delving into Discrete Normalizing Flows on SO(3) Manifold for Probabilistic Rotation Modeling
Yulin Liu*,
Haoran Liu*,
Yingda Yin*,
Yang Wang,
Baoquan Chen†,
He Wang†
CVPR 2023
arXiv /
project page /
code /
video
We propose a discrete normalizing flow on SO(3) manifold, through which one can not only effectively express arbitrary distributions on SO(3),
but also conditionally build the target distribution given input observations.
|
|
A Laplace-inspired Distribution on SO(3) for Probabilistic Rotation Estimation
Yingda Yin,
Yang Wang,
He Wang†,
Baoquan Chen†,
ICLR 2023 (Spotlight Presentation, Top 8%)
arXiv / project page /
code
We propose a novel Rotation Laplace distribution for probabilistic rotation estimation.
Rotation Laplace distribution is robust to the disturbance of outliers and enforces much gradient to the low-error region,
resulting in new state-of-the-art performance over both probabilistic and non-probabilistic baselines.
|
|
Towards Accurate Active Camera Localization
Qihang Fang*,
Yingda Yin*,
Qingnan Fan†,
Fei Xia,
Siyan Dong,
Sheng Wang,
Jue Wang,
Leonidas Guibas,
Baoquan Chen†
* Equal contribution; ordered alphabetically.
ECCV 2022
arXiv / code /
video / poster
We explicitly model the camera and scene uncertainty components to solve the problem of active camera localization by reinforcement learning.
Our algorithm improves over the state-of-the-art Markov Localization based approaches by a large margin on the fine-scale camera pose accuracy.
|
|
FisherMatch: Semi-Supervised Rotation Regression via Entropy-based Filtering
Yingda Yin,
Yingcheng Cai,
He Wang†,
Baoquan Chen†,
CVPR 2022 (Oral Presentation, Top 4%)
arXiv / project page /
code / video /
poster
We propose the first general framework for semi-supervised rotation regression.
Our algorithm models the uncertainty of rotation regression with the help of probabilistic modeling of SO(3),
which facilitates the pseudo-label filtering in semi-supervised learning.
|
|
Projective Manifold Gradient Layer for Deep Rotation Regression
Jiayi Chen,
Yingda Yin,
Tolga Birdal,
Baoquan Chen,
Leonidas Guibas,
He Wang†
CVPR 2022
arXiv / project page /
code / video /
poster
We focus on improving the backward pass of deep rotation regression. Leveraging Riemannian optimization,
we propose a SO(3) manifold-aware gradient that directly backpropagates into deep network weights.
Our plug-and-play gradient layer can be applied to different rotation representations.
|
|
Deep Reflection Prior
Yingda Yin*,
Qingnan Fan*,
Dongdong Chen,
Yujie Wang,
Angelica Aviles-Rivero,
Ruoteng Li,
Carola-Bibiane Schönlieb
Dani Lischinski,
Baoquan Chen
ArXiv 2019
arXiv
We propose a learning-based approach that captures the reflection statistical prior for single image reflection removal.
Our algorithm is driven by optimizing the target with joint constraints enhanced between multiple input images during the training stage,
but is able to eliminate reflections only from a single input for evaluation.
|
|
Peking University
Ph.D. student in Computer Applied Technology
Advisor: Prof. Baoquan Chen
2019.9 - Present
|
|
Dalian University of Technology
B.E. in Information and Communication Engineering, Innovation Class
2015.9 - 2019.7
Grades: 95.40 (Rank 1/197)
2016 & 2017 & 2018 China National Scholarships
|
|