no code implementations • 24 Jan 2025 • Huayi Zhou, Ruixiang Wang, Yunxin Tai, Yueci Deng, Guiliang Liu, Kui Jia
Bimanual robotic manipulation is a long-standing challenge of embodied intelligence due to its characteristics of dual-arm spatial-temporal coordination and high-dimensional action spaces.
no code implementations • 24 Dec 2024 • Yushu Li, Yongyi Su, Adam Goodge, Kui Jia, Xun Xu
Our method dynamically constructs a graph over text prompts, few-shot examples, and test samples, using label propagation for inference without task-specific tuning.
no code implementations • 12 Nov 2024 • Zhihao Liang, Hongdong Li, Kui Jia, Kailing Guo, Qi Zhang
Recovering the intrinsic physical attributes of a scene from images, generally termed as the inverse rendering problem, has been a central and challenging task in computer vision and computer graphics.
no code implementations • 7 Oct 2024 • Yongyi Su, Yushu Li, Nanqing Liu, Kui Jia, Xulei Yang, Chuan-Sheng Foo, Xun Xu
We then propose an effective and realistic attack method that better produces poisoned samples without access to benign samples, and derive an effective in-distribution attack objective.
1 code implementation • 26 Jul 2024 • Longkun Zou, Wanru Zhu, Ke Chen, Lihua Guo, Kailing Guo, Kui Jia, YaoWei Wang
Semantic pattern of an object point cloud is determined by its topological configuration of local geometries.
1 code implementation • 26 Jun 2024 • Yuan Gao, Yajing Luo, Junhong Wang, Kui Jia, Gui-Song Xia
Motivated by this, we propose a novel 3D generalizable relative pose estimation method by elaborating (i) with a 2. 5D shape from an RGB-D reference, (ii) with an off-the-shelf differentiable renderer, and (iii) with semantic cues from a pretrained model like DINOv2.
1 code implementation • 29 May 2024 • Yushu Li, Yongyi Su, Xulei Yang, Kui Jia, Xun Xu
In this work, we are motivated by a pitfall of TTA, i. e. sensitivity to hyper-parameters, and propose to approach HILTTA by synergizing active learning and model selection.
1 code implementation • 3 Apr 2024 • Huayi Zhou, Fei Jiang, Jin Yuan, Yong Rui, Hongtao Lu, Kui Jia
To alleviate it, we propose the first semi-supervised unconstrained head pose estimation method SemiUHPE, which can leverage abundant easily available unlabeled head images.
no code implementations • 19 Mar 2024 • Yongwei Chen, Tengfei Wang, Tong Wu, Xingang Pan, Kui Jia, Ziwei Liu
Though promising results have been achieved in single object generation, these methods often struggle to model complex 3D assets that inherently contain multiple objects.
no code implementations • 17 Mar 2024 • Zhihao Liang, Qi Zhang, WenBo Hu, Ying Feng, Lei Zhu, Kui Jia
This is because 3DGS treats each pixel as an isolated, single point rather than as an area, causing insensitivity to changes in the footprints of pixels.
no code implementations • CVPR 2024 • Bingyan Liu, Chengyu Wang, Tingfeng Cao, Kui Jia, Jun Huang
Deep Text-to-Image Synthesis (TIS) models such as Stable Diffusion have recently gained significant popularity for creative Text-to-image generation.
no code implementations • 8 Jan 2024 • Zhangjin Huang, Zhihao Liang, Haojie Zhang, Yangkai Lin, Kui Jia
Technically, we learn two parallel streams of an implicit signed distance field and an explicit surrogate surface Sur2f mesh, and unify volume rendering of the implicit signed distance function (SDF) and surface rendering of the surrogate mesh with a shared, neural shader; the unified shading promotes their convergence to the same, underlying surface.
1 code implementation • CVPR 2024 • Haojie Zhang, Yongyi Su, Xun Xu, Kui Jia
The success of large language models has inspired the computer vision community to explore image segmentation foundation model that is able to zero/few-shot generalize through prompt engineering.
1 code implementation • CVPR 2024 • Jiehong Lin, Lihua Liu, Dekun Lu, Kui Jia
Zero-shot 6D object pose estimation involves the detection of novel objects with their 6D poses in cluttered scenes, presenting significant challenges for model generalizability.
1 code implementation • CVPR 2024 • Zhihao Liang, Qi Zhang, Ying Feng, Ying Shan, Kui Jia
We propose GS-IR, a novel inverse rendering approach based on 3D Gaussian Splatting (GS) that leverages forward mapping volume rendering to achieve photorealistic novel view synthesis and relighting results.
1 code implementation • 26 Sep 2023 • Yongyi Su, Xun Xu, Kui Jia
Test-Time Adaptation aims to adapt source domain model to testing data at inference stage with success demonstrated in adapting to unseen corruptions.
no code implementations • 11 Sep 2023 • Chengyu Wang, Zhongjie Duan, Bingyan Liu, Xinyi Zou, Cen Chen, Kui Jia, Jun Huang
Text-to-image synthesis for the Chinese language poses unique challenges due to its large vocabulary size, and intricate character relationships.
1 code implementation • ICCV 2023 • Jiehong Lin, Zewei Wei, Yabin Zhang, Kui Jia
We apply the proposed VI-Net to the challenging task of category-level 6D object pose estimation for predicting the poses of unknown objects without available CAD models; experiments on the benchmarking datasets confirm the efficacy of our method, which outperforms the existing ones with a large margin in the regime of high precision.
1 code implementation • ICCV 2023 • Yushu Li, Xun Xu, Yongyi Su, Kui Jia
Existing approaches often focus on improving test-time training performance under well-curated target domain data.
1 code implementation • 30 May 2023 • KangJun Liu, Ke Chen, Lihua Guo, YaoWei Wang, Kui Jia
Inspired by good robustness of alternative dropout strategies against over-fitting on limited patterns of training samples, this paper introduces a novel concept of ShuffleMix -- Shuffle of Mixed hidden features, which can be interpreted as a kind of dropout operation in feature space.
1 code implementation • 30 May 2023 • KangJun Liu, Ke Chen, Kui Jia, YaoWei Wang
Deep representation learning is a subfield of machine learning that focuses on learning meaningful and useful representations of data through deep neural networks.
1 code implementation • 18 May 2023 • Bin Deng, Kui Jia
We hope that our investigation and the proposed simple framework can serve as a strong baseline to facilitate future studies in this field.
1 code implementation • 18 May 2023 • Yichen Zhang, Jiehong Lin, Ke Chen, Zelin Xu, YaoWei Wang, Kui Jia
Domain gap between synthetic and real data in visual regression (e. g. 6D pose estimation) is bridged in this paper via global feature alignment and local refinement on the coarse classification of discretized anchor classes in target space, which imposes a piece-wise target manifold regularization into domain-invariant representation learning.
no code implementations • 31 Mar 2023 • Yijin Chen, Xun Xu, Yongyi Su, Kui Jia
This motivates us to explore adapting an object detection model at test-time, a. k. a.
3 code implementations • ICCV 2023 • Rui Chen, Yongwei Chen, Ningxin Jiao, Kui Jia
Key to Fantasia3D is the disentangled modeling and learning of geometry and appearance.
Ranked #4 on
Text to 3D
on T$^3$Bench
1 code implementation • 20 Mar 2023 • Yongyi Su, Xun Xu, Tianrui Li, Kui Jia
Test-time training (TTT) emerges as a solution to this adaptation under a realistic scenario where access to full source domain data is not available, and instant inference on the target domain is required.
1 code implementation • CVPR 2023 • Hui Tang, Kui Jia
Moreover, we use the simulation-to-reality adaptation as a downstream task for comparing the transferability between synthetic and real data when used for pre-training, which demonstrates that synthetic data pre-training is also promising to improve real test results.
1 code implementation • CVPR 2023 • Zhihao Liang, Zhangjin Huang, Changxing Ding, Kui Jia
Recovery of an underlying scene geometry from multiview images stands as a long-time challenge in computer vision research.
1 code implementation • 23 Feb 2023 • Hui Tang, YaoWei Wang, Kui Jia
Differently, motivated by the fundamental assumption for domain adaptability, we re-cast the domain adaptation problem as discriminative clustering of target data, given strong privileged information provided by the closely related, labeled source data.
Ranked #1 on
Unsupervised Domain Adaptation
on VisDA2017
(Average Accuracy metric)
no code implementations • 30 Jan 2023 • Yabin Zhang, Bin Deng, Ruihuang Li, Kui Jia, Lei Zhang
By updating the model against the adversarial statistics perturbation during training, we allow the model to explore the worst-case domain and hence improve its generalization performance.
no code implementations • CVPR 2023 • Jiabao Lei, Jiapeng Tang, Kui Jia
More specifically, we maintain an intermediate surface mesh used for rendering new RGBD views, which subsequently becomes complete by an inpainting network; each rendered RGBD view is later back-projected as a partial surface and is supplemented into the intermediate mesh.
1 code implementation • 13 Nov 2022 • Yabin Zhang, Jiehong Lin, Ruihuang Li, Kui Jia, Lei Zhang
We also validate the effectiveness of affine transformation corruption with the Transformer backbones, where we decompose the reconstruction of the complete point cloud into the reconstructions of detailed local patches and rough global shape, alleviating the position leakage problem in the reconstruction.
1 code implementation • 20 Oct 2022 • Yongwei Chen, Rui Chen, Jiabao Lei, Yabin Zhang, Kui Jia
Creation of 3D content by stylization is a promising yet challenging problem in computer vision and graphics research.
1 code implementation • 11 Oct 2022 • Hongyang Li, Jiehong Lin, Kui Jia
Establishment of point correspondence between camera and object coordinate systems is a promising way to solve 6D object poses.
1 code implementation • 16 Aug 2022 • Bin Deng, Kui Jia
First, we show that the key assumption of support overlap of invariant features used in IB-IRM is strong for the guarantee of OOD generalization and it is still possible to achieve the optimal solution without this assumption.
1 code implementation • 3 Aug 2022 • KangJun Liu, Ke Chen, Kui Jia
Such target coding schemes are less flexible to model inter-class correlation and are sensitive to sparse and imbalanced data distribution as well.
1 code implementation • 12 Jul 2022 • Jiehong Lin, Zewei Wei, Changxing Ding, Kui Jia
It is difficult to precisely annotate object instances and their semantics in 3D space, and as such, synthetic data are extensively used for these tasks, e. g., category-level 6D object pose and size estimation.
1 code implementation • 7 Jul 2022 • Wentao Tan, Changxing Ding, Pengfei Wang, Mingming Gong, Kui Jia
This common practice causes the model to overfit to existing feature styles in the source domain, resulting in sub-optimal generalization ability on target domains.
1 code implementation • 7 Jul 2022 • Yabin Zhang, Jiehong Lin, Chenhang He, Yongwei Chen, Kui Jia, Lei Zhang
In this work, we make the first attempt, to the best of our knowledge, to consider the local geometry information explicitly into the masked auto-encoding, and propose a novel Masked Surfel Prediction (MaskSurf) method.
1 code implementation • 6 Jun 2022 • Yongyi Su, Xun Xu, Kui Jia
Test-time training (TTT) emerges as a solution to this adaptation under a realistic scenario where access to full source domain data is not available and instant inference on target domain is required.
1 code implementation • 16 May 2022 • Jinpeng Lin, Zhihao Liang, Shengheng Deng, Lile Cai, Tao Jiang, Tianrui Li, Kui Jia, Xun Xu
We demonstrate the effectiveness of the proposed method on the nuScenes dataset and show that it outperforms existing AL strategies significantly.
1 code implementation • 7 May 2022 • Zelin Xu, Yichen Zhang, Ke Chen, Kui Jia
Inspired by the success of point-pair features, the goal of this paper is to recover the 6D pose of an object instance segmented from RGB-D images by locally matching pairs of oriented points between the model and camera space.
no code implementations • 6 May 2022 • Yongyi Su, Xun Xu, Kui Jia
Addressing the annotation challenge in 3D Point Cloud segmentation has inspired research into weakly supervised learning.
no code implementations • 5 May 2022 • Zhangjin Huang, Yuxin Wen, ZiHao Wang, Jinjuan Ren, Kui Jia
For example, while deep learning methods are increasingly popular, our systematic studies suggest that, surprisingly, a few classical methods perform even better in terms of both robustness and generalization; our studies also suggest that the practical challenges of misalignment of point sets from multi-view scanning, missing of surface points, and point outliers remain unsolved by all the existing surface reconstruction methods.
no code implementations • 2 May 2022 • Xian Shi, Xun Xu, Wanyue Zhang, Xiatian Zhu, Chuan Sheng Foo, Kui Jia
We also demonstrate the feasibility of a more efficient training strategy.
1 code implementation • CVPR 2022 • Shengheng Deng, Zhihao Liang, Lin Sun, Kui Jia
These multi-view methods either refine the proposals predicted from single view via fused features, or fuse the features without considering the global spatial context; their performance is limited consequently.
2 code implementations • CVPR 2022 • Yabin Zhang, Minghan Li, Ruihuang Li, Kui Jia, Lei Zhang
In this work, we, for the first time to our best knowledge, propose to perform Exact Feature Distribution Matching (EFDM) by exactly matching the empirical Cumulative Distribution Functions (eCDFs) of image features, which could be implemented by applying the Exact Histogram Matching (EHM) in the image feature space.
Ranked #7 on
Style Transfer
on StyleBench
1 code implementation • 8 Mar 2022 • Yongwei Chen, ZiHao Wang, Longkun Zou, Ke Chen, Kui Jia
Such a challenge of Simulation-to-Reality (Sim2Real) domain gap could be mitigated via learning algorithms of domain adaptation; however, we argue that generation of synthetic point clouds via more physically realistic rendering is a powerful alternative, as systematic non-uniform noise patterns can be captured.
no code implementations • CVPR 2022 • Hui Tang, Kui Jia
To answer these problems, we propose a novel Taylor expansion inspired filtration (TEIF) framework, which admits the samples of moderate confidence with similar feature or gradient to the respective one averaged over the labeled and highly confident unlabeled data.
1 code implementation • NeurIPS 2021 • Jiehong Lin, Hongyang Li, Ke Chen, Jiangbo Lu, Kui Jia
In this paper, we propose a novel design of Sparse Steerable Convolution (SS-Conv) to address the shortcoming; SS-Conv greatly accelerates steerable convolution with sparse tensors, while strictly preserving the property of SE(3)-equivariance.
1 code implementation • 22 Aug 2021 • Pengfei Wang, Changxing Ding, Wentao Tan, Mingming Gong, Kui Jia, DaCheng Tao
In particular, the performance of our unsupervised UCF method in the MSMT17$\to$Market1501 task is better than that of the fully supervised setting on Market1501.
1 code implementation • 20 Aug 2021 • Longkun Zou, Hui Tang, Ke Chen, Kui Jia
The point cloud representation of an object can have a large geometric variation in view of inconsistent data acquisition procedure, which thus leads to domain discrepancy due to diverse and uncontrollable shape representation cross datasets.
1 code implementation • ICCV 2021 • Zhihao Liang, Zhihao LI, Songcen Xu, Mingkui Tan, Kui Jia
State-of-the-art methods largely rely on a general pipeline that first learns point-wise features discriminative at semantic and instance levels, followed by a separate step of point grouping for proposing object instances.
Ranked #10 on
3D Instance Segmentation
on S3DIS
1 code implementation • 30 Jun 2021 • Yong Guo, Yaofo Chen, Mingkui Tan, Kui Jia, Jian Chen, Jingdong Wang
In practice, the convolutional operation on some of the windows (e. g., smooth windows that contain very similar pixels) can be very redundant and may introduce noises into the computation.
1 code implementation • ICCV 2021 • Mingkui Tan, Zhuangwei Zhuang, Sitao Chen, Rong Li, Kui Jia, Qicheng Wang, Yuanqing Li
We then explore more efficient contextual modules under perspective projection and fuse the LiDAR features into the camera stream to boost the performance of the two-stream network.
Ranked #9 on
Semantic Segmentation
on KITTI-360
no code implementations • 18 Jun 2021 • Yabin Zhang, Bin Deng, Kui Jia, Lei Zhang
Domain adaptation becomes more challenging with increasing gaps between source and target domains.
1 code implementation • 18 Jun 2021 • Jiabao Lei, Kui Jia, Yi Ma
More specifically, we identify from the linear regions, partitioned by an MLP based implicit function, the analytic cells and analytic faces that are associated with the function's zero-level isosurface.
no code implementations • 1 Jun 2021 • Yabin Zhang, Haojian Zhang, Bin Deng, Shuai Li, Kui Jia, Lei Zhang
Especially, state-of-the-art SSL methods significantly outperform existing UDA methods on the challenging UDA benchmark of DomainNet, and state-of-the-art UDA methods could be further enhanced with SSL techniques.
1 code implementation • ICCV 2021 • Jiapeng Tang, Jiabao Lei, Dan Xu, Feiying Ma, Kui Jia, Lei Zhang
To this end, we propose to learn implicit surface reconstruction by sign-agnostic optimization of convolutional occupancy networks, to simultaneously achieve advanced scalability to large-scale scenes, generality to novel shapes, and applicability to raw scans in a unified framework.
1 code implementation • 10 Apr 2021 • Bin Deng, Yabin Zhang, Hui Tang, Changxing Ding, Kui Jia
The great promise that UB$^2$DA makes, however, brings significant learning challenges, since domain adaptation can only rely on the predictions of unlabeled target data in a partially overlapped label space, by accessing the interface of source model.
1 code implementation • CVPR 2021 • Jiapeng Tang, Dan Xu, Kui Jia, Lei Zhang
This paper focuses on the task of 4D shape reconstruction from a sequence of point clouds.
1 code implementation • CVPR 2021 • Shengheng Deng, Xun Xu, Chaozheng Wu, Ke Chen, Kui Jia
The ability to understand the ways to interact with objects from visual cues, a. k. a.
Ranked #1 on
Affordance Detection
on 3D AffordanceNet
1 code implementation • ICCV 2021 • Jiehong Lin, Zewei Wei, Zhihao LI, Songcen Xu, Kui Jia, Yuanqing Li
DualPoseNet stacks two parallel pose decoders on top of a shared pose encoder, where the implicit decoder predicts object poses with a working mechanism different from that of the explicit one; they thus impose complementary supervision on the training of pose encoder.
Ranked #4 on
6D Pose Estimation using RGBD
on REAL275
1 code implementation • 5 Mar 2021 • Hui Tang, Kui Jia
Unsupervised domain adaptation aims to learn a task classifier that performs well on the unlabeled target domain, by utilizing the labeled source domain.
no code implementations • 18 Jan 2021 • Xian Shi, Xun Xu, Ke Chen, Lile Cai, Chuan Sheng Foo, Kui Jia
Deep learning models are the state-of-the-art methods for semantic point cloud segmentation, the success of which relies on the availability of large-scale annotated datasets.
1 code implementation • 8 Jan 2021 • Haojian Zhang, Yabin Zhang, Kui Jia, Lei Zhang
Unsupervised domain adaptation (UDA) aims to learn models for a target domain of unlabeled data by transferring knowledge from a labeled source domain.
1 code implementation • ICCV 2021 • Longkun Zou, Hui Tang, Ke Chen, Kui Jia
The point cloud representation of an object can have a large geometric variation in view of inconsistent data acquisition procedure, which thus leads to domain discrepancy due to diverse and uncontrollable shape representation cross datasets.
no code implementations • 18 Dec 2020 • Zelin Xu, Ke Chen, KangJun Liu, Changxing Ding, YaoWei Wang, Kui Jia
By adapting existing ModelNet40 and ScanNet datasets to the single-view, partial setting, experiment results can verify the necessity of object pose estimation and superiority of our PAPNet to existing classifiers.
no code implementations • 14 Dec 2020 • Chaozheng Wu, Lin Sun, Xun Xu, Kui Jia
Given the large shape variations among different instances of a same category, we are formally interested in developing a quantity defined for individual points on a continuous object surface; the quantity specifies how individual surface points contribute to the formation of the shape as the category.
no code implementations • CVPR 2021 • Wenbin Zhao, Jiabao Lei, Yuxin Wen, JianGuo Zhang, Kui Jia
Motivated from a universal phenomenon that self-similar shape patterns of local surface patches repeat across the entire surface of an object, we aim to push forward the data-driven strategies and propose to learn a local implicit surface network for a shared, adaptive modeling of the entire surface for a direct surface reconstruction from raw point cloud; we also enhance the leveraging of surface self-similarities by improving correlations among the optimized latent codes of individual surface patches.
no code implementations • CVPR 2021 • Mingyue Yang, Yuxin Wen, Weikai Chen, Yongwei Chen, Kui Jia
Many learning-based approaches have difficulty scaling to unseen data, as the generality of its learned prior is limited to the scale and variations of the training samples.
2 code implementations • 8 Dec 2020 • Hui Tang, Xiatian Zhu, Ke Chen, Kui Jia, C. L. Philip Chen
To address this issue, we are motivated by a UDA assumption of structural similarity across domains, and propose to directly uncover the intrinsic target discrimination via constrained clustering, where we constrain the clustering solutions using structural source regularization that hinges on the very same assumption.
no code implementations • 16 Nov 2020 • Jinmiao Cai, Nianjuan Jiang, Xiaoguang Han, Kui Jia, Jiangbo Lu
Skeleton-based action recognition has attracted research attentions in recent years.
no code implementations • ICML 2020 • Yuxin Wen, Shuai Li, Kui Jia
However, it is observed that such methods would lead to standard performance degradation, i. e., the degradation on natural examples.
no code implementations • 9 Oct 2020 • Yao Li, Xianggang Yu, Xiaoguang Han, Nianjuan Jiang, Kui Jia, Jiangbo Lu
In this work, we propose an interactive system to design diverse high-quality garment images from fashion sketches and the texture information.
2 code implementations • NeurIPS 2020 • Chaozheng Wu, Jian Chen, Qiaoyu Cao, Jianchi Zhang, Yunxin Tai, Lin Sun, Kui Jia
To test GPNet, we contribute a synthetic dataset of 6-DOF object grasps; evaluation is conducted using rule-based criteria, simulation test, and real test.
1 code implementation • 10 Sep 2020 • Jiehong Lin, Xian Shi, Yuan Gao, Ke Chen, Kui Jia
Point set is arguably the most direct approximation of an object or scene surface, yet its practical acquisition often suffers from the shortcoming of being noisy, sparse, and possibly incomplete, which restricts its use for a high-quality surface recovery.
1 code implementation • 13 Aug 2020 • Jiapeng Tang, Xiaoguang Han, Mingkui Tan, Xin Tong, Kui Jia
However, they all have their own drawbacks, and cannot properly reconstruct the surface shapes of complex topologies, arguably due to a lack of constraints on the topologicalstructures in their learning frameworks.
1 code implementation • ECCV 2020 • Yabin Zhang, Bin Deng, Kui Jia, Lei Zhang
To make the proposed A$^2$LP useful for UDA, we propose empirical schemes to generate such virtual instances.
1 code implementation • CVPR 2020 • Yuan Gao, Haoping Bai, Zequn Jie, Jiayi Ma, Kui Jia, Wei Liu
We propose to incorporate neural architecture search (NAS) into general-purpose multi-task learning (GP-MTL).
2 code implementations • CVPR 2020 • Hui Tang, Ke Chen, Kui Jia
To alleviate this risk, we are motivated by the assumption of structural domain similarity, and propose to directly uncover the intrinsic target discrimination via discriminative clustering of target data.
1 code implementation • 10 Mar 2020 • Kun Zhou, Xiaoguang Han, Nianjuan Jiang, Kui Jia, Jiangbo Lu
Estimating 3D human pose from a single image is a challenging task.
2 code implementations • 20 Feb 2020 • Yabin Zhang, Bin Deng, Hui Tang, Lei Zhang, Kui Jia
By using MCSD as a measure of domain distance, we develop a new domain adaptation bound for multi-class UDA; its data-dependent, probably approximately correct bound is also developed that naturally suggests adversarial learning objectives to align conditional feature distributions across source and target domains.
1 code implementation • ICML 2020 • Jiabao Lei, Kui Jia
This paper studies a problem of learning surface mesh via implicit functions in an emerging field of deep learning surface reconstruction, where implicit functions are popularly implemented as multi-layer perceptrons (MLPs) with rectified linear units (ReLU).
no code implementations • 14 Jan 2020 • Lulu Tang, Ke Chen, Chaozheng Wu, Yu Hong, Kui Jia, Zhi-Xin Yang
Existing deep learning algorithms for point cloud analysis mainly concern discovering semantic patterns from global configuration of local geometries in a supervised learning manner.
no code implementations • ECCV 2020 • Qi Chen, Lin Sun, Zhixin Wang, Kui Jia, Alan Yuille
Accurate 3D object detection in LiDAR based point clouds suffers from the challenges of data sparsity and irregularities.
Ranked #3 on
3D Object Detection
on KITTI Pedestrians Moderate
no code implementations • 26 Dec 2019 • Zelin Xu, Ke Chen, Kui Jia
Solving 6D pose estimation is non-trivial to cope with intrinsic appearance and shape variation and severe inter-object occlusion, and is made more challenging in light of extrinsic large illumination changes and low quality of the acquired data under an uncontrolled environment.
Ranked #1 on
6D Pose Estimation using RGBD
on LineMOD
(Mean ADD-S metric)
1 code implementation • 24 Dec 2019 • Huanglin Yu, Ke Chen, Kaiqi Wang, Yanlin Qian, Zhao-Xiang Zhang, Kui Jia
Regressing the illumination of a scene from the representations of object appearances is popularly adopted in computational color constancy.
2 code implementations • 24 Dec 2019 • Yuxin Wen, Jiehong Lin, Ke Chen, C. L. Philip Chen, Kui Jia
Regularizing the targeted attack loss with our proposed geometry-aware objectives results in our proposed method, Geometry-Aware Adversarial Attack ($GeoA^3$).
2 code implementations • 27 Nov 2019 • Hui Tang, Kui Jia
Based on an integrated category and domain classifier, DADA has a novel adversarial objective that encourages a mutually inhibitory relation between category and domain predictions for any input instance.
Ranked #1 on
Synthetic-to-Real Translation
on Syn2Real-C
3 code implementations • NeurIPS 2019 • Jiezhang Cao, Langyuan Mo, Yifan Zhang, Kui Jia, Chunhua Shen, Mingkui Tan
Multiple marginal matching problem aims at learning mappings to match a source domain to multiple target domains and it has attracted great attention in many applications, such as multi-domain image translation.
no code implementations • ICCV 2019 • Kun Zhou, Xiaoguang Han, Nianjuan Jiang, Kui Jia, Jiangbo Lu
Estimating 3D human pose from a single image is a challenging task.
Ranked #1 on
Monocular 3D Human Pose Estimation
on Human3.6M
(Use Video Sequence metric, using extra
training data)
no code implementations • 25 Sep 2019 • Yuxin Wen, Jiehong Lin, Ke Chen, Kui Jia
Recent studies show that machine learning models are vulnerable to adversarial examples.
1 code implementation • ICCV 2019 • Ziyang Wu, Yuwei Li, Lihua Guo, Kui Jia
However, due to the inherent local connectivity of CNN, the CNN-based relation network (RN) can be sensitive to the spatial position relationship of semantic objects in two compared images.
1 code implementation • ICCV 2019 • Junyi Pan, Xiaoguang Han, Weikai Chen, Jiapeng Tang, Kui Jia
The key to our approach is a novel progressive shaping framework that alternates between mesh deformation and topology modification.
Ranked #3 on
3D Shape Reconstruction
on Pix3D
1 code implementation • 15 May 2019 • Kui Jia, Shuai Li, Yuxin Wen, Tongliang Liu, DaCheng Tao
To this end, we first prove that DNNs are of local isometry on data distributions of practical interest; by using a new covering of the sample space and introducing the local isometry property of DNNs into generalization analysis, we establish a new generalization error bound that is both scale- and range-sensitive to singular value spectrum of each of networks' weight matrices.
no code implementations • 25 Apr 2019 • Kui Jia, Jiehong Lin, Mingkui Tan, DaCheng Tao
Such a perspective enables us to study deep multi-view learning in the context of regularized network training, for which we present control experiments of benchmark image classification to show the efficacy of our proposed CorrReg.
1 code implementation • 23 Apr 2019 • Bolun Cai, Xiangmin Xu, Xiaofen Xing, Kui Jia, Jie Miao, DaCheng Tao
Visual tracking is challenging due to image variations caused by various factors, such as object deformation, scale change, illumination change and occlusion.
1 code implementation • CVPR 2019 2019 • Jiapeng Tang, Xiaoguang Han, Junyi Pan, Kui Jia, Xin Tong
To this end, we propose in this paper a skeleton-bridged, stage-wise learning approach to address the challenge.
1 code implementation • CVPR 2019 • Yabin Zhang, Hui Tang, Kui Jia, Mingkui Tan
Since target samples are unlabeled, we also propose a scheme of cross-domain training to help learn the target classifier.
1 code implementation • CVPR 2019 • Jiapeng Tang, Xiaoguang Han, Junyi Pan, Kui Jia, Xin Tong
To this end, we propose in this paper a skeleton-bridged, stage-wise learning approach to address the challenge.
2 code implementations • 5 Mar 2019 • Zhixin Wang, Kui Jia
We also propose component variants of F-ConvNet, including an FCN variant that extracts multi-resolution frustum features, and a refined use of F-ConvNet over a reduced 3D space.
Ranked #3 on
3D Object Detection
on KITTI Cyclists Easy
no code implementations • 12 Feb 2019 • Chaorui Deng, Qi Wu, Guanghui Xu, Zhuliang Yu, Yanwu Xu, Kui Jia, Mingkui Tan
Most state-of-the-art methods in VG operate in a two-stage manner, wherein the first stage an object detector is adopted to generate a set of object proposals from the input image and the second stage is simply formulated as a cross-modal matching problem that finds the best match between the language query and all region proposals.
no code implementations • 25 Dec 2018 • Lin Sun, Kui Jia, Yuejia Shen, Silvio Savarese, Dit Yan Yeung, Bertram E. Shi
To learn from these heterogenous input sources, existing methods reply on two-stream architectural designs that contain independent, parallel streams of Recurrent Neural Networks (RNNs).
Action Recognition In Videos
Multi-Person Pose Estimation
+2
no code implementations • 27 Sep 2018 • Shuai Li, Kui Jia
We present a formal measure-theoretical theory of neural networks (NN) built on {\it probability coupling theory}.
no code implementations • 21 Sep 2018 • Kun Zhou, Jinmiao Cai, Yao Li, Yulong Shi, Xiaoguang Han, Nianjuan Jiang, Kui Jia, Jiangbo Lu
In this paper, a novel deep-learning based framework is proposed to infer 3D human poses from a single image.
2 code implementations • ECCV 2018 • Yabin Zhang, Hui Tang, Kui Jia
Fine-grained visual categorization (FGVC) is challenging due in part to the fact that it is often difficult to acquire an enough number of training samples.
Ranked #1 on
Fine-Grained Image Classification
on DIB-10K
(using extra training data)
no code implementations • 25 Jun 2018 • Yulong Shi, Xiaoguang Han, Nianjuan Jiang, Kun Zhou, Kui Jia, Jiangbo Lu
Although significant advances have been made in the area of human poses estimation from images using deep Convolutional Neural Network (ConvNet), it remains a big challenge to perform 3D pose inference in-the-wild.
Ranked #245 on
3D Human Pose Estimation
on Human3.6M
2 code implementations • 16 Jun 2018 • Yabin Zhang, Kui Jia, Zhixin Wang
In this work, we propose a Weakly Supervised Part Detection Network (PartNet) that is able to detect discriminative local parts for use of fine-grained categorization.
no code implementations • 4 Dec 2017 • Bolun Cai, Xiangmin Xu, Kailing Guo, Kui Jia, DaCheng Tao
With the powerful down-sampling process, the co-training DSN set a new state-of-the-art performance for image super-resolution.
no code implementations • ICCV 2017 • Bolun Cai, Xianming Xu, Kailing Guo, Kui Jia, Bin Hu, DaCheng Tao
We propose a joint intrinsic-extrinsic prior model to estimate both illumination and reflectance from an observed image.
no code implementations • ICCV 2017 • Lin Sun, Kui Jia, Kevin Chen, Dit Yan Yeung, Bertram E. Shi, Silvio Savarese
This method effectively enhances the ability to model dynamics across time and addresses the non-stationary issue of long-term motion dynamics without significantly increasing the model complexity.
no code implementations • 30 Dec 2016 • Shuai Li, Kui Jia, Xiaogang Wang
The recent successful deep neural networks are largely trained in a supervised manner.
no code implementations • CVPR 2017 • Kui Jia
To this end, we propose to constrain the solutions of weight matrices in the orthogonal feasible set during the whole process of network training, and achieve this by a simple yet effective method called Singular Value Bounding (SVB).
3 code implementations • 28 Jan 2016 • Bolun Cai, Xiangmin Xu, Kui Jia, Chunmei Qing, DaCheng Tao
The key to achieve haze removal is to estimate a medium transmission map for an input hazy image.
Ranked #7 on
Image Dehazing
on RS-Haze
no code implementations • 4 Jan 2016 • Abrar H. Abdulnabi, Gang Wang, Jiwen Lu, Kui Jia
Each CNN will generate attribute-specific feature representations, and then we apply multi-task learning on the features to predict their attributes.
no code implementations • ICCV 2015 • Lin Sun, Kui Jia, Dit-yan Yeung, Bertram E. Shi
Human actions in video sequences are three-dimensional (3D) spatio-temporal signals characterizing both the visual appearance and motion dynamics of the involved humans and objects.
no code implementations • 20 Jan 2015 • Yuting Zhang, Kui Jia, Yueming Wang, Gang Pan, Tsung-Han Chan, Yi Ma
By assuming a human face as piece-wise planar surfaces, where each surface corresponds to a facial part, we develop in this paper a Constrained Part-based Alignment (CPA) algorithm for face recognition across pose and/or expression.
no code implementations • CVPR 2014 • Lin Sun, Kui Jia, Tsung-Han Chan, Yuqiang Fang, Gang Wang, Shuicheng Yan
In this paper, we propose to combine SFA with deep learning techniques to learn hierarchical representations from the video data itself.
no code implementations • CVPR 2014 • Tianzhu Zhang, Kui Jia, Changsheng Xu, Yi Ma, Narendra Ahuja
The proposed part matching tracker (PMT) has a number of attractive properties.
2 code implementations • 14 Apr 2014 • Tsung-Han Chan, Kui Jia, Shenghua Gao, Jiwen Lu, Zinan Zeng, Yi Ma
In this work, we propose a very simple deep learning network for image classification which comprises only the very basic data processing components: cascaded principal component analysis (PCA), binary hashing, and block-wise histograms.
Ranked #40 on
Image Classification
on MNIST
no code implementations • 31 Mar 2014 • Kui Jia, Tsung-Han Chan, Zinan Zeng, Shenghua Gao, Gang Wang, Tianzhu Zhang, Yi Ma
The task is to identify the inlier features and establish their consistent correspondences across the image set.
no code implementations • CVPR 2013 • Zinan Zeng, Shijie Xiao, Kui Jia, Tsung-Han Chan, Shenghua Gao, Dong Xu, Yi Ma
Our framework is motivated by the observation that samples from the same class repetitively appear in the collection of ambiguously labeled training images, while they are just ambiguously labeled in each image.