Feng Wen
Research Subjects
Feng Wen has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Aixue Ye
Huawei Noahs Ark Lab, Huawei Technologies Company Ltd., Beijing, China
3 shared publications - Shaoyi Du
Xi’an Jiaotong University
3 shared publications - Zhiqiang Tian
School of Software Engineering, Xian Jiaotong University, Xian, China
3 shared publications - Hongbo Zhang
Noah's Ark Lab Huawei
2 shared publications - Runzhao Yao
College of Artificial Intelligence, Institute of Robotics and Automatic Information System, Nankai University, Tianjin, China
2 shared publications - Wenting Cui
Xi’an Jiaotong University
2 shared publications
Latest Publications
- 23 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (1) , 657-674
Translating Images to Road Network: A Sequence-to-Sequence Perspective
The extraction of road network is essential for the generation of high-definition maps since it enables the precise localization of road landmarks and their interconnections. However, generating road network poses a significant challenge due to the conflicting underlying combination of Euclidean (e.g., road landmarks location) and non-Euclidean (e.g., road topological connectivity) structures. Existing methods struggle to merge the two types of data domains effectively, but few of them address it properly. Instead, our work establishes a unified representation of both types of data domain by projecting both Euclidean and non-Euclidean data into an integer series called RoadNet Sequence. Further than modeling an auto-regressive sequence-to-sequence Transformer model to understand RoadNet Sequence, we decouple the dependency of RoadNet Sequence into a mixture of auto-regressive and non-autoregressive dependency. Building on this, our proposed non-autoregressive sequence-to-sequence approach leverages non-autoregressive dependencies while fixing the gap towards auto-regressive dependencies, resulting in success in both efficiency and accuracy. We further identify two main bottlenecks in the current RoadNetTransformer on a non-overfitting split of the dataset: poor landmark detection limited by the BEV Encoder and error propagation to topology reasoning. Therefore, we propose Topology-Inherited Training to inherit better topology knowledge into RoadNetTransformer. Additionally, we collect SD-Maps from open-source map datasets and use this prior information to significantly improve landmark detection and reachability. Extensive experiments on the nuScenes dataset demonstrate the superiority of RoadNet Sequence representation and the non-autoregressive approach compared to existing state-of-the-art alternatives.
- 23 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (1) , 657-674
Translating Images to Road Network: A Sequence-to-Sequence Perspective
The extraction of road network is essential for the generation of high-definition maps since it enables the precise localization of road landmarks and their interconnections. However, generating road network poses a significant challenge due to the conflicting underlying combination of Euclidean (e.g., road landmarks location) and non-Euclidean (e.g., road topological connectivity) structures. Existing methods struggle to merge the two types of data domains effectively, but few of them address it properly. Instead, our work establishes a unified representation of both types of data domain by projecting both Euclidean and non-Euclidean data into an integer series called RoadNet Sequence. Further than modeling an auto-regressive sequence-to-sequence Transformer model to understand RoadNet Sequence, we decouple the dependency of RoadNet Sequence into a mixture of auto-regressive and non-autoregressive dependency. Building on this, our proposed non-autoregressive sequence-to-sequence approach leverages non-autoregressive dependencies while fixing the gap towards auto-regressive dependencies, resulting in success in both efficiency and accuracy. We further identify two main bottlenecks in the current RoadNetTransformer on a non-overfitting split of the dataset: poor landmark detection limited by the BEV Encoder and error propagation to topology reasoning. Therefore, we propose Topology-Inherited Training to inherit better topology knowledge into RoadNetTransformer. Additionally, we collect SD-Maps from open-source map datasets and use this prior information to significantly improve landmark detection and reachability. Extensive experiments on the nuScenes dataset demonstrate the superiority of RoadNet Sequence representation and the non-autoregressive approach compared to existing state-of-the-art alternatives.
- 19 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 46 (9) , 6154-6166
Generative Variational-Contrastive Learning for Self-Supervised Point Cloud Representation
Self-supervised representation learning for 3D point clouds has attracted increasing attention. However, existing methods in the field of 3D computer vision generally use fixed embeddings to represent the latent features, and impose hard constraints on the embeddings to make the latent feature values of the positive samples converge to consistency, which limits the ability of feature extractors to generalize over different data domains. To address this issue, we propose a Generative Variational-Contrastive Learning (GVC) model, where Gaussian distribution is used to construct a continuous, smoothed representation of the latent features. A distribution constraint and cross-supervision are constructed to improve the transfer ability of the feature extractor over synthetic and real-world data. Specifically, we design a variational contrastive module to constrain the feature distribution instead of feature values corresponding to each sample in the latent space. Moreover, a generative cross-supervision module is introduced to preserve the invariance features and promote the consistency of feature distribution among positive samples. Experimental results demonstrate that GVC achieves SOTA on different downstream tasks. In particular, with only pre-training on the synthetic dataset, GVC achieves a lead of 8.4% and 14.2% when transferring to the real-world dataset in the linear classification and few-shot classification.
- 19 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 46 (9) , 6154-6166
Generative Variational-Contrastive Learning for Self-Supervised Point Cloud Representation
Self-supervised representation learning for 3D point clouds has attracted increasing attention. However, existing methods in the field of 3D computer vision generally use fixed embeddings to represent the latent features, and impose hard constraints on the embeddings to make the latent feature values of the positive samples converge to consistency, which limits the ability of feature extractors to generalize over different data domains. To address this issue, we propose a Generative Variational-Contrastive Learning (GVC) model, where Gaussian distribution is used to construct a continuous, smoothed representation of the latent features. A distribution constraint and cross-supervision are constructed to improve the transfer ability of the feature extractor over synthetic and real-world data. Specifically, we design a variational contrastive module to constrain the feature distribution instead of feature values corresponding to each sample in the latent space. Moreover, a generative cross-supervision module is introduced to preserve the invariance features and promote the consistency of feature distribution among positive samples. Experimental results demonstrate that GVC achieves SOTA on different downstream tasks. In particular, with only pre-training on the synthetic dataset, GVC achieves a lead of 8.4% and 14.2% when transferring to the real-world dataset in the linear classification and few-shot classification.
- 30 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 900-913
RDD: Learning Reinforced 3D Detectors and Descriptors Based on Policy Gradient
Keypoint detection and descriptor matching are two vital steps in the 3D feature extraction framework, but they are difficult to learn in an end-to-end fashion due to their inherent discreteness. To tackle the non-differentiable operations, we formulate feature extraction as a decision-making problem: the network is treated as a policy pool that can make probabilistic estimations for keypoint selection and feature matching, supervised by maximizing a reward expectation of actions. In this way, we propose a novel end-to-end training paradigm of 3D feature extraction based on the stochastic policy gradient method, named Reinforced Detectors and Descriptors (RDD). Firstly, we propose a local-to-global probabilistic keypoint selection module that formulates the sampling probabilities of keypoints in a local-and-global mechanism to yield sparse and accurate keypoints. Secondly, we regard feature matching as an optimal transport problem and an efficient Sinkhorn method is leveraged to solve the optimal matching probabilities. In particular, we carefully design a reward function and derive gradients of probabilistic actions, thus overcoming the discreteness and providing reinforced supervision signals. Since our reward function is calculated from sampled keypoints rather than from randomly sampled points as in existing methods, the gap between training and inference is bridged. Experimental results demonstrate that our approach exceeds the quality of state-of-the-art methods and shows strong generalization ability. Remarkably, our approach can achieve significantly higher Registration Recall than other advanced methods when aligning scenes with a small number of keypoints, due to our highly accurate and repeatable detector.
- 30 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 900-913
RDD: Learning Reinforced 3D Detectors and Descriptors Based on Policy Gradient
Keypoint detection and descriptor matching are two vital steps in the 3D feature extraction framework, but they are difficult to learn in an end-to-end fashion due to their inherent discreteness. To tackle the non-differentiable operations, we formulate feature extraction as a decision-making problem: the network is treated as a policy pool that can make probabilistic estimations for keypoint selection and feature matching, supervised by maximizing a reward expectation of actions. In this way, we propose a novel end-to-end training paradigm of 3D feature extraction based on the stochastic policy gradient method, named Reinforced Detectors and Descriptors (RDD). Firstly, we propose a local-to-global probabilistic keypoint selection module that formulates the sampling probabilities of keypoints in a local-and-global mechanism to yield sparse and accurate keypoints. Secondly, we regard feature matching as an optimal transport problem and an efficient Sinkhorn method is leveraged to solve the optimal matching probabilities. In particular, we carefully design a reward function and derive gradients of probabilistic actions, thus overcoming the discreteness and providing reinforced supervision signals. Since our reward function is calculated from sampled keypoints rather than from randomly sampled points as in existing methods, the gap between training and inference is bridged. Experimental results demonstrate that our approach exceeds the quality of state-of-the-art methods and shows strong generalization ability. Remarkably, our approach can achieve significantly higher Registration Recall than other advanced methods when aligning scenes with a small number of keypoints, due to our highly accurate and repeatable detector.
- 11 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (12) , 14760-14776
Hunter: Exploring High-Order Consistency for Point Cloud Registration With Severe Outliers
After decades of investigation, point cloud registration is still a challenging task in practice, especially when the correspondences are contaminated by a large number of outliers. It may result in a rapidly decreasing probability of generating a hypothesis close to the true transformation, leading to the failure of point cloud registration. To tackle this problem, we propose a transformation estimation method, named Hunter, for robust point cloud registration with severe outliers. The core of Hunter is to design a global-to-local exploration scheme to robustly find the correct correspondences. The global exploration aims to exploit guided sampling to generate promising initial alignments. To this end, a hypergraph-based consistency reasoning module is introduced to learn the high-order consistency among correct correspondences, which is able to yield a more distinct inlier cluster that facilitates the generation of all-inlier hypotheses. Moreover, we propose a preference-based local exploration module that exploits the preference information of top-$k$ promising hypotheses to find a better transformation. This module can efficiently obtain multiple reliable transformation hypotheses by using a multi-initialization searching strategy. Finally, we present a distance-angle based hypothesis selection criterion to choose the most reliable transformation, which can avoid selecting symmetrically aligned false transformations. Experimental results on simulated, indoor, and outdoor datasets, demonstrate that Hunter can achieve significant superiority over the state-of-the-art methods, including both learning-based and traditional methods (as shown in Fig. 1). Moreover, experimental results also indicate that Hunter can achieve more stable performance compared with all other methods with severe outliers.- 11 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (12) , 14760-14776
Hunter: Exploring High-Order Consistency for Point Cloud Registration With Severe Outliers
After decades of investigation, point cloud registration is still a challenging task in practice, especially when the correspondences are contaminated by a large number of outliers. It may result in a rapidly decreasing probability of generating a hypothesis close to the true transformation, leading to the failure of point cloud registration. To tackle this problem, we propose a transformation estimation method, named Hunter, for robust point cloud registration with severe outliers. The core of Hunter is to design a global-to-local exploration scheme to robustly find the correct correspondences. The global exploration aims to exploit guided sampling to generate promising initial alignments. To this end, a hypergraph-based consistency reasoning module is introduced to learn the high-order consistency among correct correspondences, which is able to yield a more distinct inlier cluster that facilitates the generation of all-inlier hypotheses. Moreover, we propose a preference-based local exploration module that exploits the preference information of top-$k$ promising hypotheses to find a better transformation. This module can efficiently obtain multiple reliable transformation hypotheses by using a multi-initialization searching strategy. Finally, we present a distance-angle based hypothesis selection criterion to choose the most reliable transformation, which can avoid selecting symmetrically aligned false transformations. Experimental results on simulated, indoor, and outdoor datasets, demonstrate that Hunter can achieve significant superiority over the state-of-the-art methods, including both learning-based and traditional methods (as shown in Fig. 1). Moreover, experimental results also indicate that Hunter can achieve more stable performance compared with all other methods with severe outliers.- 29 July 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53 (21) , 24804-24821
Relative order constraint for monocular depth estimation
- 29 July 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53 (21) , 24804-24821
Relative order constraint for monocular depth estimation
- Open Access
- 12 August 2022
- Published by Frontiers Media SA in Frontiers in Earth Science
- Vol. 10
Domain Adaptation on Multiple Cloud Recognition From Different Types of Meteorological Satellite
Meteorological satellites have become an indispensable meteorological tool for earth observation, as aiding in areas such as cloud detection, which has important guiding significance for maritime activities. However, it is time-consuming and labor-intensive to obtain fine-grained annotations provided by artificial experience or mature satellite cloud products for multi-spectral maritime cloud imageries, especially when new satellites are launched. Moreover, due to the data discrepancy caused by different detection bands, existing models have inadequate generalization performance compared to new satellites, and some cannot be directly migrated. In this paper, to reduce the data distribution’s discrepancy, an approach is presented based on unsupervised domain adaption method for marine cloud detection task based on Himawari-8 satellite data as a source domain and Fengyun-4 satellite data as a target domain. The goal of the proposed method is to leverage the representation power of adversarial learning to extract domain-invariant features, consisting of a segmentation model, a feature extract model for target domain, and a domain discriminator. In addition, aiming to remedy the discrepancy of detection bands, a band mapping module is designed to implement consistency between different bands. The result of the experiments demonstrated the effectiveness of the proposed method with a 7% improvement compared with the comparative experiment. We also designed a series of statistical experiments on different satellite data to further study cloudy perception representation, including data visualization experiment and cloud type statistics.
- 12 August 2022
- Published by Frontiers Media SA in Frontiers in Earth Science
- Vol. 10
Open AccessDomain Adaptation on Multiple Cloud Recognition From Different Types of Meteorological Satellite
Meteorological satellites have become an indispensable meteorological tool for earth observation, as aiding in areas such as cloud detection, which has important guiding significance for maritime activities. However, it is time-consuming and labor-intensive to obtain fine-grained annotations provided by artificial experience or mature satellite cloud products for multi-spectral maritime cloud imageries, especially when new satellites are launched. Moreover, due to the data discrepancy caused by different detection bands, existing models have inadequate generalization performance compared to new satellites, and some cannot be directly migrated. In this paper, to reduce the data distribution’s discrepancy, an approach is presented based on unsupervised domain adaption method for marine cloud detection task based on Himawari-8 satellite data as a source domain and Fengyun-4 satellite data as a target domain. The goal of the proposed method is to leverage the representation power of adversarial learning to extract domain-invariant features, consisting of a segmentation model, a feature extract model for target domain, and a domain discriminator. In addition, aiming to remedy the discrepancy of detection bands, a band mapping module is designed to implement consistency between different bands. The result of the experiments demonstrated the effectiveness of the proposed method with a 7% improvement compared with the comparative experiment. We also designed a series of statistical experiments on different satellite data to further study cloudy perception representation, including data visualization experiment and cloud type statistics.
- 27 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 5470-5475
BSP-MonoLoc: Basic Semantic Primitives based Monocular Localization on Roads
Robust visual localization in traffic scenes is a fundamental problem for self-driving vehicles. However, it is still challenging to achieve accurate localization performance because of drastic viewpoint and illumination changes. To address the issues, we design a novel monocular localization framework based on a light-weight prior map, called BSP-MonoLoc, which leverages the 2D semantic primitives from the monocular images and the 3D basic semantic primitives from the prior map. These primitives are commonly available but lack of distinctive signature. To effectively make associations between the 2D and 3D primitives and refine the vehicle’s pose, we adopt an iterative optimization method, where an efficient hierarchical sample strategy is designed to give a good initial prediction for the associations and the pose. Experimental results on the KAIST dataset and our dataset demonstrate the proposed method can achieve high localization accuracy and run at a real-time performance.
- 27 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 5470-5475
BSP-MonoLoc: Basic Semantic Primitives based Monocular Localization on Roads
Robust visual localization in traffic scenes is a fundamental problem for self-driving vehicles. However, it is still challenging to achieve accurate localization performance because of drastic viewpoint and illumination changes. To address the issues, we design a novel monocular localization framework based on a light-weight prior map, called BSP-MonoLoc, which leverages the 2D semantic primitives from the monocular images and the 3D basic semantic primitives from the prior map. These primitives are commonly available but lack of distinctive signature. To effectively make associations between the 2D and 3D primitives and refine the vehicle’s pose, we adopt an iterative optimization method, where an efficient hierarchical sample strategy is designed to give a good initial prediction for the associations and the pose. Experimental results on the KAIST dataset and our dataset demonstrate the proposed method can achieve high localization accuracy and run at a real-time performance.
- 1 December 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 234-240
Design and Implementation of a Heterogeneous Relational Database Synchronization Mechanism Based on Tree Distribution Architecture
In allusion to the database server of tree distribution in the network environment, this paper proposes a heterogeneous database synchronization mechanism based on the shadow table method. In this paper, the application environment of the isolation netgap is described, and the overall architecture of the synchronization system is introduced, and the implementation of the module such as change data capture, incremental file generation, data synchronization operation and exception handling is proposed. Finally, the database synchronization mechanism is verified by the application of the data acquisition system of an organization.
- 1 December 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 234-240
Design and Implementation of a Heterogeneous Relational Database Synchronization Mechanism Based on Tree Distribution Architecture
In allusion to the database server of tree distribution in the network environment, this paper proposes a heterogeneous database synchronization mechanism based on the shadow table method. In this paper, the application environment of the isolation netgap is described, and the overall architecture of the synchronization system is introduced, and the implementation of the module such as change data capture, incremental file generation, data synchronization operation and exception handling is proposed. Finally, the database synchronization mechanism is verified by the application of the data acquisition system of an organization.
- Open Access
- 1 December 2019
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1419 (1) , 012015
Evolution of urban construction land structure in Henan Province based on information entropy and equilibrium model
The paper aims to accurately identify the evolution characteristics of urban land use structure in Henan Province, clarify the current situation and problems of urban land use structure, and provide reference and basis for the optimization of urban land use structure and urban land allocation. Statistical analysis, comparative analysis, information entropy, equilibrium and dominance were used. From 2007 to 2016, the urban construction land area in Henan Province showed an overall upward trend; from 2007 to 2016, the per capita construction land area in Henan Province showed an overall upward - downward - upward trend. From 2007 to 2016, the information entropy, equilibrium and dominance curves of urban land use structure in Henan Province are in a stable state, which shows that the change of urban land use structure tends to be gentle, the overall fluctuation range is not large, and all types of land are relatively uniform. To a certain extent, it can show that urban land development in Henan Province is relatively mature. In the structure of urban construction land in Henan Province, the proportion of residential land is the highest, and its area and proportion are increasing year by year. This is closely related to the increase of urban population. The proportion of industrial land is higher, but it decreases year by year. The land for urban green space, square and road traffic facilities is increasing year by year, and others. Although the land use area of different types is increasing year by year, the proportion of structure is decreasing year by year, and the urban land use structure of Henan Province is in the stage of continuous adjustment.
- 1 December 2019
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1419 (1) , 012015
Open AccessEvolution of urban construction land structure in Henan Province based on information entropy and equilibrium model
The paper aims to accurately identify the evolution characteristics of urban land use structure in Henan Province, clarify the current situation and problems of urban land use structure, and provide reference and basis for the optimization of urban land use structure and urban land allocation. Statistical analysis, comparative analysis, information entropy, equilibrium and dominance were used. From 2007 to 2016, the urban construction land area in Henan Province showed an overall upward trend; from 2007 to 2016, the per capita construction land area in Henan Province showed an overall upward - downward - upward trend. From 2007 to 2016, the information entropy, equilibrium and dominance curves of urban land use structure in Henan Province are in a stable state, which shows that the change of urban land use structure tends to be gentle, the overall fluctuation range is not large, and all types of land are relatively uniform. To a certain extent, it can show that urban land development in Henan Province is relatively mature. In the structure of urban construction land in Henan Province, the proportion of residential land is the highest, and its area and proportion are increasing year by year. This is closely related to the increase of urban population. The proportion of industrial land is higher, but it decreases year by year. The land for urban green space, square and road traffic facilities is increasing year by year, and others. Although the land use area of different types is increasing year by year, the proportion of structure is decreasing year by year, and the urban land use structure of Henan Province is in the stage of continuous adjustment.
- Open Access
- 1 January 2014
- Published by Elsevier in Acta Physico-Chimica Sinica
- Vol. 30 (8) , 1501-1508
Self-Assembly of Vesicles from Oligoaramide Based on Multiple Hydrogen Bonds
- 1 January 2014
- Published by Elsevier in Acta Physico-Chimica Sinica
- Vol. 30 (8) , 1501-1508
Open AccessSelf-Assembly of Vesicles from Oligoaramide Based on Multiple Hydrogen Bonds