Yaxin Peng

Shanghai University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Chaomin Shen

East China Normal University

Shihui Ying

Shanghai University

Guixu Zhang

East China Normal University

Jian Tang

Midea Group (China)

Zhengping Che

Midea Group (China)

Tieyong Zeng

Chinese University of Hong Kong

Yichen Zhu

Midea Group (China)

Zhiyuan Xu

Midea Group (China)

Feifei Feng

Midea Group (China)

Yangchun Zhang

Shanghai University

Cooperative Institutions

Shanghai University

Chinese Academy of Sciences

East China Normal University

Shanghai Jiao Tong University

Huazhong University of Science and Technology

Midea Group (China)

Tsinghua University

Dalian University of Technology

University of Chinese Academy of Sciences

University of Electronic Science and Technology of China

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Variational Color Image Segmentation via Chromaticity-Brightness Decomposition

Lecture notes in computer science (2009)

Zheng Bao Yajun Liu Yaxin Peng Guixu Zhang

Chromaticity

Lightness

10.1007/978-3-642-11301-7_31

Cite

Citations (7)

A novel statistical method for 3D range data registration based on Lie group framework

Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE (2012)

Yaxin Peng Wei Lin Chaomin Shen Shihui Ying

Registration of 3D range data is to find the transformation that best maps one data set to the other. In this paper, Lie group parametric representation is combined with the Expectation Maximization (EM) method to provide a unified framework. First, having a transformation fixed, the EM algorithm is introduced to find the correspondence between two data sets through correspondence probability, which covers the relationship of all points, instead of using exact correspondence such as the classical Iterative Closest Point (ICP) method. With this type of ststistical correspondence, we could deal with the presence of the degradations such as outliers and incomplete point sets. Second, having the updated correspondence fixed, and introducing Lie group parametric representation, the transformation is updated by minimizing a quadratic programming. Then, an alternative iterative strategy by the above two steps is used to approximate the desired correspondence and transformation. The comparative experiment between our Lie-EM-ICP algorithm and Lie-ICP algorithm using point cloud is presented. Our algorithm is demonstrated to be accurate and robust, especially in the presence of incomplete point sets and outliers.

Iterative closest point

10.1117/12.977288

Cite

Citations (0)

Exploring prototype-guided strategy for domain decomposition in physics-informed neural network

Nonlinear Dynamics (2025)

Yanjie Wang Yaxin Peng Zhaoping Hu Ying Li

10.1007/s11071-025-10871-4

Cite

Citations (0)

Variational-based speckle noise removal of SAR imagery

Chaomin Shen Ling Pi Yaxin Peng Zhibin Li

In this paper we present a variational method for synthetic aperture radar (SAR) speckle removal. Variational method is a newly developed technique for the removal of SAR's multiplicative noise. For an image, we could define an energy functional. The energy evolves as the original image changes, and the minimum energy corresponds to the speckle reduced result. Partial differential equation (PDE) technique is used to get the minimal solution. Our energy functional makes use of the statistical information of the multiplicative noise since it follows a Gamma law with mean mu = 1 and variance sigma ² = 1/M for M-look SAR. Our energy is a regularization term with two constraints. The regularization term is the integral for the norm of image gradient; two constraints are the mean of noise should be 1 and the variance of noise should be 1/M. We use the method of Lagrange multipliers, Euler-Lagrange equation and heat flow method to obtain the minimizer of the energy. ERS Precision Image (PRI) data are to demonstrate our algorithm. Numerical result shows that the speckle reduced image preserves edges and point targets while smoothes homogenous regions in the original image. The algorithm is computationally efficient and easy to implement.

Regularization

Multiplicative noise

Speckle noise

Energy functional

10.1109/igarss.2007.4422848

Cite

Citations (2)

A Coarse-to-Fine Generalized-ICP Algorithm With Trimmed Strategy

IEEE Access (2020)

Xin Wang Yun Li Yaxin Peng Shihui Ying

In this paper, we introduce a modified Generalized Iterative Closest Point (GICP) algorithm by presenting a coarse-to-fine strategy. Our contributions can be summarized as: Firstly, we use adaptively a plane-to-plane probabilistic matching model by gradually reducing the neighborhood range for given two point sets. It is an inner coarse-to-fine iteration process. Secondly, we use an outer coarse-to-fine strategy to bridge the point-to-point and plane-to-plane registration for refining the matching. Thirdly, we use the trimmed method to gradually eliminate the effects of incorrect correspondences, which improves the robustness of the methods especially for the low overlap cases. Moreover, we also extend our method to the scale registration case. Finally, we conduct extensive experiments to demonstrate that our method is more reliable and robust in various situations, including missing points, noise and different scale factors. Experimental results show that our approach outperforms several state-of-the-art registration methods.

Robustness

Iterative closest point

Point set registration

10.1109/access.2020.2976132

Cite

Citations (17)

Hybrid Atlas Building with Deep Registration Priors

arXiv (Cornell University) (2021)

Nian Wu Jian Wang Miaomiao Zhang Guixu Zhang Yaxin Peng

Registration-based atlas building often poses computational challenges in high-dimensional image spaces. In this paper, we introduce a novel hybrid atlas building algorithm that fast estimates atlas from large-scale image datasets with much reduced computational cost. In contrast to previous approaches that iteratively perform registration tasks between an estimated atlas and individual images, we propose to use learned priors of registration from pre-trained neural networks. This newly developed hybrid framework features several advantages of (i) providing an efficient way of atlas building without losing the quality of results, and (ii) offering flexibility in utilizing a wide variety of deep learning based registration methods. We demonstrate the effectiveness of this proposed model on 3D brain magnetic resonance imaging (MRI) scans.

Image registration

Deep Neural Networks

Brain atlas

10.48550/arxiv.2112.06406

Cite

Citations (0)

VAN: Voting and Attention Based Network for Unsupervised Medical Image Registration

Lecture notes in computer science (2021)

Zhiang Zu Guixu Zhang Yaxin Peng Zhen Ye Chaomin Shen

Normalization

Image registration

10.1007/978-3-030-89188-6_29

Cite

Citations (0)

APFN: Adaptive Perspective-based Fusion Network for 3D Place Recognition

IEEE Transactions on Instrumentation and Measurement (2024)

Zhu Jian-xiang Keni Yang Yangchun Zhang Yan Peng Yaxin Peng

Sensor Fusion

10.1109/tim.2024.3418083

Cite

Citations (0)

A new initialization method based on normed statistical spaces in deep networks

Inverse Problems and Imaging (2020)

Hongfei Yang Xiaofeng Ding Raymond Chan Hui Hu Yaxin Peng

Training deep neural networks can be difficult. For classical neural networks, the initialization method by Xavier and Yoshua which is later generalized by He, Zhang, Ren and Sun can facilitate stable training. However, with the recent development of new layer types, we find that the above mentioned initialization methods may fail to lead to successful training. Based on these two methods, we will propose a new initialization by studying the parameter space of a network. Our principal is to put constrains on the growth of parameters in different layers in a consistent way. In order to do so, we introduce a norm to the parameter space and use this norm to measure the growth of parameters. Our new method is suitable for a wide range of layer types, especially for layers with parameter-sharing weight matrices.

Initialization

10.3934/ipi.2020045

Cite

Citations (7)

Recognizable Information Bottleneck

arXiv (Cornell University) (2023)

Yilin Lyu Xin Liu Mingyang Song Xinyue Wang Yaxin Peng

Information Bottlenecks (IBs) learn representations that generalize to unseen data by information compression. However, existing IBs are practically unable to guarantee generalization in real-world scenarios due to the vacuous generalization bound. The recent PAC-Bayes IB uses information complexity instead of information compression to establish a connection with the mutual information generalization bound. However, it requires the computation of expensive second-order curvature, which hinders its practical application. In this paper, we establish the connection between the recognizability of representations and the recent functional conditional mutual information (f-CMI) generalization bound, which is significantly easier to estimate. On this basis we propose a Recognizable Information Bottleneck (RIB) which regularizes the recognizability of representations through a recognizability critic optimized by density ratio matching under the Bregman divergence. Extensive experiments on several commonly used datasets demonstrate the effectiveness of the proposed method in regularizing the model and estimating the generalization gap.

Information bottleneck method

Cartographic generalization

Divergence (linguistics)

Basis (linear algebra)

10.48550/arxiv.2304.14618

Cite

Citations (0)