HHHaifeng Huang

Haifeng Huang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Haifeng Huang has a total of 65 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    RoboGround: Robotic Manipulation with Grounded Vision-Language Priors

    by Haifeng Huang, Xinyi Chen, Yilun Chen, Hao Li, Xiaoshen Han, Zehan Wang, Tai Wang, Jiangmiao Pang, Zhou Zhao

    Recent advancements in robotic manipulation have highlighted the potential of intermediate representations for improving policy generalization. In this work, we explore grounding masks as an effective intermediate representation, balancing two key advantages: (1) effective spatial guidance that specifies target objects and placement areas while also conveying information about object shape and size, and (2) broad generalization potential driven by large-scale vision-language models pretrained on diverse grounding datasets. We introduce RoboGround, a grounding-aware robotic manipulation system that leverages grounding masks as an intermediate representation to guide policy networks in object manipulation tasks. To further explore and enhance generalization, we propose an automated pipeline for generating large-scale, simulated data with a diverse set of objects and instructions. Extensive experiments show the value of our dataset and the effectiveness of grounding masks as intermediate guidance, significantly enhancing the generalization abilities of robot policies.

  • Open Access

    Multi-Modal Domain Adaptation Across Video Scenes for Temporal Video Grounding

    by Haifeng Huang, Yang Zhao, Zehan Wang, Yan Xia, Zhou Zhao

    Temporal Video Grounding (TVG) aims to localize the temporal boundary of a specific segment in an untrimmed video based on a given language query. Since datasets in this domain are often gathered from limited video scenes, models tend to overfit to scene-specific factors, which leads to suboptimal performance when encountering new scenes in real-world applications. In a new scene, the fine-grained annotations are often insufficient due to the expensive labor cost, while the coarse-grained video-query pairs are easier to obtain. Thus, to address this issue and enhance model performance on new scenes, we explore the TVG task in an unsupervised domain adaptation (UDA) setting across scenes for the first time, where the video-query pairs in the source scene (domain) are labeled with temporal boundaries, while those in the target scene are not. Under the UDA setting, we introduce a novel Adversarial Multi-modal Domain Adaptation (AMDA) method to adaptively adjust the model's scene-related knowledge by incorporating insights from the target data. Specifically, we tackle the domain gap by utilizing domain discriminators, which help identify valuable scene-related features effective across both domains. Concurrently, we mitigate the semantic gap between different modalities by aligning video-query pairs with related semantics. Furthermore, we employ a mask-reconstruction approach to enhance the understanding of temporal semantics within a scene. Extensive experiments on Charades-STA, ActivityNet Captions, and YouCook2 demonstrate the effectiveness of our proposed method.

  • Open Access

    Distilling Coarse-to-Fine Semantic Matching Knowledge for Weakly Supervised 3D Visual Grounding

    by Zehan Wang, Haifeng Huang, Yang Zhao, Linjun Li, Xize Cheng, Yichen Zhu, Aoxiong Yin, Zhou Zhao

    3D visual grounding involves finding a target object in a 3D scene that corresponds to a given sentence query. Although many approaches have been proposed and achieved impressive performance, they all require dense object-sentence pair annotations in 3D point clouds, which are both time-consuming and expensive. To address the problem that fine-grained annotated data is difficult to obtain, we propose to leverage weakly supervised annotations to learn the 3D visual grounding model, i.e., only coarse scene-sentence correspondences are used to learn object-sentence links. To accomplish this, we design a novel semantic matching model that analyzes the semantic similarity between object proposals and sentences in a coarse-to-fine manner. Specifically, we first extract object proposals and coarsely select the top-K candidates based on feature and class similarity matrices. Next, we reconstruct the masked keywords of the sentence using each candidate one by one, and the reconstructed accuracy finely reflects the semantic similarity of each candidate to the query. Additionally, we distill the coarse-to-fine semantic matching knowledge into a typical two-stage 3D visual grounding model, which reduces inference costs and improves performance by taking full advantage of the well-studied structure of the existing architectures. We conduct extensive experiments on ScanRefer, Nr3D, and Sr3D, which demonstrate the effectiveness of our proposed method.


  • Open Access

    3DRP-Net: 3D Relative Position-aware Network for 3D Visual Grounding

    by Zehan Wang, Haifeng Huang, Yang Zhao, Linjun Li, Xize Cheng, Yichen Zhu, Aoxiong Yin, Zhou Zhao

    3D visual grounding aims to localize the target object in a 3D point cloud by a free-form language description. Typically, the sentences describing the target object tend to provide information about its relative relation between other objects and its position within the whole scene. In this work, we propose a relation-aware one-stage framework, named 3D Relative Position-aware Network (3DRP-Net), which can effectively capture the relative spatial relationships between objects and enhance object attributes. Specifically, 1) we propose a 3D Relative Position Multi-head Attention (3DRP-MA) module to analyze relative relations from different directions in the context of object pairs, which helps the model to focus on the specific object relations mentioned in the sentence. 2) We designed a soft-labeling strategy to alleviate the spatial ambiguity caused by redundant points, which further stabilizes and enhances the learning process through a constant and discriminative distribution. Extensive experiments conducted on three benchmarks (i.e., ScanRefer and Nr3D/Sr3D) demonstrate that our method outperforms all the state-of-the-art methods in general. The source code will be released on GitHub.

  • Open Access

    Distilling Coarse-to-Fine Semantic Matching Knowledge for Weakly Supervised 3D Visual Grounding

    by Zehan Wang, Haifeng Huang, Yang Zhao, Linjun Li, Xize Cheng, Yichen Zhu, Aoxiong Yin, Zhou Zhao

    3D visual grounding involves finding a target object in a 3D scene that corresponds to a given sentence query. Although many approaches have been proposed and achieved impressive performance, they all require dense object-sentence pair annotations in 3D point clouds, which are both time-consuming and expensive. To address the problem that fine-grained annotated data is difficult to obtain, we propose to leverage weakly supervised annotations to learn the 3D visual grounding model, i.e., only coarse scene-sentence correspondences are used to learn object-sentence links. To accomplish this, we design a novel semantic matching model that analyzes the semantic similarity between object proposals and sentences in a coarse-to-fine manner. Specifically, we first extract object proposals and coarsely select the top-K candidates based on feature and class similarity matrices. Next, we reconstruct the masked keywords of the sentence using each candidate one by one, and the reconstructed accuracy finely reflects the semantic similarity of each candidate to the query. Additionally, we distill the coarse-to-fine semantic matching knowledge into a typical two-stage 3D visual grounding model, which reduces inference costs and improves performance by taking full advantage of the well-studied structure of the existing architectures. We conduct extensive experiments on ScanRefer, Nr3D, and Sr3D, which demonstrate the effectiveness of our proposed method.

  • Simulation research on the whole process of spaceborne XTI-SAR oceanic eddy detection

    by Yuhao Lu, Xiaoqing Wang, Qingsong Wang, Haifeng Huang

    Oceanic eddy is a highly occurring marine phenomenon in the ocean, characterized by closed circulation. Oceanic eddy can cause changes in Sea Surface Height (SSH) and will make the SSH around the center of the oceanic eddy to be higher or lower than the surrounding area, so the identification of oceanic eddy can be achieved by detecting the change of SSH. Traditionally, radar altimeter is an important means to measure the sea surface height, but its resolution is low, so it can’t effectively detect sub-mesoscale oceanic eddies. With its all-time, all-weather, wide-swath and high-resolution characteristics, Synthetic Aperture Radar (SAR) has become a research hotspot in oceanic eddy remote sensing detection. It uses radar antennas from different angles of view to form a baseline, called interferometric SAR, which can achieve the measurement of SSH. In this paper, a signal level simulation model of oceanic eddy for spaceborne cross-track interferometric SAR(XTI-SAR) is established, and an evaluation method based on ideal interferometric factors is introduced. In the experiments, we analyze the influence on the error of oceanic eddy SSH measurement at different sea states. The results show that the increase of wind speed will increase the height measurement error of oceanic eddy, and the difference of wind direction will cause obvious changes in the height measurement error.

  • Open Access

    3DRP-Net: 3D Relative Position-aware Network for 3D Visual Grounding

    by Zehan Wang, Haifeng Huang, Yang Zhao, Linjun Li, Xize Cheng, Yichen Zhu, Aoxiong Yin, Zhou Zhao

    Zehan Wang, Haifeng Huang, Yang Zhao, Linjun Li, Xize Cheng, Yichen Zhu, Aoxiong Yin, Zhou Zhao. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.


  • Composite energetic salt based on 3-nitramino-4-(1H-tetrazol-5-yl)furazan

    by Kun Zhao, Yameng Shi, Hongli Li, Haifeng Huang, Jun Yang


  • Open Access

    Mg Doped Li–LiB Alloy with In Situ Formed Lithiophilic LiB Skeleton for Lithium Metal Batteries

    by Chen Wu, Haifeng Huang, Weiyi Lu, Zengxi Wei, Xuyan Ni, Fu Sun, Piao Qing, Zhijian Liu, Jianmin Ma, Weifeng Wei, et al.

    High energy density lithium metal batteries (LMBs) are promising next‐generation energy storage devices. However, the uncontrollable dendrite growth and huge volume change limit their practical applications. Here, a new Mg doped Li–LiB alloy with in situ formed lithiophilic 3D LiB skeleton (hereinafter called Li–B–Mg composite) is presented to suppress Li dendrite and mitigate volume change. The LiB skeleton exhibits superior lithiophilic and conductive characteristics, which contributes to the reduction of the local current density and homogenization of incoming Li+ flux. With the introduction of Mg, the composite achieves an ultralong lithium deposition/dissolution lifespan (500 h, at 0.5 mA cm−2) without short circuit in the symmetrical battery. In addition, the electrochemical performance is superior in full batteries assembled with LiCoO2 cathode and the manufactured composite. The currently proposed 3D Li–B–Mg composite anode may significantly propel the advancement of LMB technology from laboratory research to industrial commercialization.


  • Research on Simulation Evaluation Method of Sea Surface Height Measurement Based on Ideal Interferometric Factors

    by Yuxin Hou, Haifeng Huang, Shuang Wu, Xiaoqing Wang, Daoxiang An, Zhimin Zhou

    The sea surface height measurement using interferometric synthetic aperture radar (InSAR) has the characteristic of high accuracy, wide swath; the motion in the sea must be taken into account. The simulation evaluation method based on ideal interferometric factors is proposed, and the calculation method of ideal interferometric factors for common bistatic interferometric Range-Doppler (RD) model is deduced. The evaluation process and complete steps are given. The simulation is carried out, using typical spaceborne InSAR parameters. The result shows that the interferometric phase of sea surface is of non-zero mean, which is in accordance with the law of elevation bias caused by the dynamic characteristics of ocean, and verifies that this method is suitable for the evaluation of sea surface height measurement.