1 code implementation • 14 Sep 2023 • Zhiwu Qing, Shiwei Zhang, Ziyuan Huang, Yingya Zhang, Changxin Gao, Deli Zhao, Nong Sang
When pre-training on the large-scale Kinetics-710, we achieve 89. 7% on Kinetics-400 with a frozen ViT-L model, which verifies the scalability of DiST.
1 code implementation • 24 Aug 2023 • Huaxin Zhang, Xiang Wang, Xiaohao Xu, Zhiwu Qing, Changxin Gao, Nong Sang
For snippet-level learning, we introduce an online-updated memory to store reliable snippet prototypes for each class.
1 code implementation • 31 Jul 2023 • Feng Zhang, Bin Xu, Zhiqiang Li, Xinran Liu, Qingbo Lu, Changxin Gao, Nong Sang
To address this issue, we introduce a new perspective to synthesize the signal-independent noise by a generative model.
Ranked #2 on
Image Denoising
on SID SonyA7S2 x300
1 code implementation • 15 May 2023 • Jialong Zuo, Changqian Yu, Nong Sang, Changxin Gao
Extensive experiments demonstrate that our model not only significantly improves existing methods on all these tasks, but also shows great ability in the few-shot and domain generalization settings.
Ranked #3 on
Text based Person Retrieval
on ICFG-PEDES
Pedestrian Attribute Recognition
Person Re-Identification
+3
1 code implementation • CVPR 2023 • Xiang Wang, Shiwei Zhang, Zhiwu Qing, Changxin Gao, Yingya Zhang, Deli Zhao, Nong Sang
To address these issues, we develop a Motion-augmented Long-short Contrastive Learning (MoLo) method that contains two crucial components, including a long-short contrastive objective and a motion autodecoder.
1 code implementation • 6 Mar 2023 • Xiang Wang, Shiwei Zhang, Jun Cen, Changxin Gao, Yingya Zhang, Deli Zhao, Nong Sang
Learning from large-scale contrastive language-image pre-training like CLIP has shown remarkable success in a wide range of downstream tasks recently, but it is still under-explored on the challenging few-shot action recognition (FSAR) task.
no code implementations • 12 Jan 2023 • Dongyue Wu, Zilin Guo, Aoyan Li, Changqian Yu, Changxin Gao, Nong Sang
Under this novel view, we propose a Class Center Similarity layer (CCS layer) to address the above-mentioned challenges by generating adaptive class centers conditioned on different scenes and supervising the similarities between class centers.
1 code implementation • 9 Jan 2023 • Xiang Wang, Shiwei Zhang, Zhiwu Qing, Zhengrong Zuo, Changxin Gao, Rong Jin, Nong Sang
To be specific, HyRSM++ consists of two key components, a hybrid relation module and a temporal set matching metric.
no code implementations • 9 Jan 2023 • Huan Peng, Fenggang Liu, Yangguang Li, Bin Huang, Jing Shao, Nong Sang, Changxin Gao
Human-Object Interaction (HOI) detection aims to learn how human interacts with surrounding objects.
no code implementations • IEEE Transactions on Pattern Analysis and Machine Intelligence 2022 • Chuchu Han, Zhedong Zheng, Kai Su, Dongdong Yu, Zehuan Yuan, Changxin Gao, Nong Sang, Yi Yang
Person search aims at localizing and recognizing query persons from raw video frames, which is a combination of two sub-tasks, i. e., pedestrian detection and person re-identification.
Ranked #3 on
Person Search
on PRW
no code implementations • 2 Nov 2022 • Yixuan Pei, Zhiwu Qing, Jun Cen, Xiang Wang, Shiwei Zhang, Yaxiong Wang, Mingqian Tang, Nong Sang, Xueming Qian
The former is to reduce the memory cost by preserving only one condensed frame instead of the whole video, while the latter aims to compensate the lost spatio-temporal details in the Frame Condensing stage.
1 code implementation • 24 Jul 2022 • Zhiwu Qing, Shiwei Zhang, Ziyuan Huang, Xiang Wang, Yuehuan Wang, Yiliang Lv, Changxin Gao, Nong Sang
Inspired by this, we propose propose Masked Action Recognition (MAR), which reduces the redundant computation by discarding a proportion of patches and operating only on a part of the videos.
Ranked #9 on
Action Recognition
on Something-Something V2
no code implementations • 18 Jun 2022 • Xiang Wang, Huaxin Zhang, Shiwei Zhang, Changxin Gao, Yuanjie Shao, Nong Sang
This technical report presents our first place winning solution for temporal action detection task in CVPR-2022 AcitivityNet Challenge.
1 code implementation • CVPR 2022 • Xiang Wang, Shiwei Zhang, Zhiwu Qing, Mingqian Tang, Zhengrong Zuo, Changxin Gao, Rong Jin, Nong Sang
To overcome the two limitations, we propose a novel Hybrid Relation guided Set Matching (HyRSM) approach that incorporates two key components: hybrid relation module and set matching metric.
no code implementations • CVPR 2022 • Zhiwu Qing, Shiwei Zhang, Ziyuan Huang, Yi Xu, Xiang Wang, Mingqian Tang, Changxin Gao, Rong Jin, Nong Sang
In this work, we aim to learn representations by leveraging more abundant information in untrimmed videos.
no code implementations • 12 Mar 2022 • Fusen Wang, Kai Liu, Fei Long, Nong Sang, Xiaofeng Xia, Jun Sang
However, the transformer directly partitions the crowd images into a series of tokens, which may not be a good choice due to each pedestrian being an independent individual, and the parameter number of the network is very large.
no code implementations • 22 Dec 2021 • Yuhang Wu, Tengteng Huang, Haotian Yao, Chi Zhang, Yuanjie Shao, Chuchu Han, Changxin Gao, Nong Sang
First, we present a Domain-Specific Contrastive Learning (DSCL) mechanism to fully explore intradomain information by comparing samples only from the same domain.
Contrastive Learning
Domain Adaptive Person Re-Identification
+2
1 code implementation • 15 Dec 2021 • Zongheng Huang, Yifan Sun, Chuchu Han, Changxin Gao, Nong Sang
By combining two fundamental learning approaches in DML, e. g., classification training and pairwise training, we set up a strong baseline for ZS-SBIR.
no code implementations • 5 Dec 2021 • Moran Li, Haibin Huang, Yi Zheng, Mengtian Li, Nong Sang, Chongyang Ma
In this work, we present a new method for 3D face reconstruction from sparse-view RGB images.
1 code implementation • 3 Dec 2021 • Feng Zhang, Yuanjie Shao, Yishi Sun, Kai Zhu, Changxin Gao, Nong Sang
We introduce a Noise Disentanglement Module (NDM) to disentangle the noise and content in the reflectance maps with the reliable aid of unpaired clean images.
Ranked #1 on
Low-Light Image Enhancement
on MEF
(NIQE metric)
1 code implementation • 25 Nov 2021 • Rui Wang, Jian Chen, Gang Yu, Li Sun, Changqian Yu, Changxin Gao, Nong Sang
Image manipulation with StyleGAN has been an increasing concern in recent years. Recent works have achieved tremendous success in analyzing several semantic latent spaces to edit the attributes of the generated images. However, due to the limited semantic and spatial manipulation precision in these latent spaces, the existing endeavors are defeated in fine-grained StyleGAN image manipulation, i. e., local attribute translation. To address this issue, we discover attribute-specific control units, which consist of multiple channels of feature maps and modulation styles.
2 code implementations • 21 Sep 2021 • Changqian Yu, Yuanjie Shao, Changxin Gao, Nong Sang
The last layer of FCN is typically a global classifier (1x1 convolution) to recognize each pixel to a semantic label.
Ranked #17 on
Semantic Segmentation
on PASCAL Context
no code implementations • ICCV 2021 • Chuchu Han, Kai Su, Dongdong Yu, Zehuan Yuan, Changxin Gao, Nong Sang, Yi Yang, Changhu Wang
Large-scale labeled training data is often difficult to collect, especially for person identities.
1 code implementation • 24 Aug 2021 • Zhiwu Qing, Ziyuan Huang, Shiwei Zhang, Mingqian Tang, Changxin Gao, Marcelo H. Ang Jr, Rong Jin, Nong Sang
The visualizations show that ParamCrop adaptively controls the center distance and the IoU between two augmented views, and the learned change in the disparity along the training process is beneficial to learning a strong representation.
no code implementations • 24 Jun 2021 • Zhiwu Qing, Xiang Wang, Ziyuan Huang, Yutong Feng, Shiwei Zhang, Jianwen Jiang, Mingqian Tang, Changxin Gao, Nong Sang
Temporal action localization aims to localize starting and ending time with action category.
1 code implementation • ICCV 2021 • Xiang Wang, Shiwei Zhang, Zhiwu Qing, Yuanjie Shao, Zhengrong Zuo, Changxin Gao, Nong Sang
Most recent approaches for online action detection tend to apply Recurrent Neural Network (RNN) to capture long-range temporal structure.
Ranked #7 on
Online Action Detection
on THUMOS'14
1 code implementation • 20 Jun 2021 • Xiang Wang, Zhiwu Qing, Ziyuan Huang, Yutong Feng, Shiwei Zhang, Jianwen Jiang, Mingqian Tang, Changxin Gao, Nong Sang
We calculate the detection results by assigning the proposals with corresponding classification results.
Ranked #1 on
Temporal Action Localization
on ActivityNet-1.3
(using extra training data)
no code implementations • 20 Jun 2021 • Xiang Wang, Zhiwu Qing, Ziyuan Huang, Yutong Feng, Shiwei Zhang, Jianwen Jiang, Mingqian Tang, Yuanjie Shao, Nong Sang
Then our proposed Local-Global Background Modeling Network (LGBM-Net) is trained to localize instances by using only video-level labels based on Multi-Instance Learning (MIL).
Weakly-supervised Learning
Weakly-supervised Temporal Action Localization
+1
1 code implementation • 13 Jun 2021 • Zhiwu Qing, Ziyuan Huang, Xiang Wang, Yutong Feng, Shiwei Zhang, Jianwen Jiang, Mingqian Tang, Changxin Gao, Marcelo H. Ang Jr, Nong Sang
This technical report analyzes an egocentric video action detection method we used in the 2021 EPIC-KITCHENS-100 competition hosted in CVPR2021 Workshop.
1 code implementation • 9 Jun 2021 • Ziyuan Huang, Zhiwu Qing, Xiang Wang, Yutong Feng, Shiwei Zhang, Jianwen Jiang, Zhurong Xia, Mingqian Tang, Nong Sang, Marcelo H. Ang Jr
In this paper, we present empirical results for training a stronger video vision transformer on the EPIC-KITCHENS-100 Action Recognition dataset.
no code implementations • 4 Jun 2021 • Fusen Wang, Jun Sang, Zhongyuan Wu, Qi Liu, Nong Sang
In this paper, we propose a Hybrid Attention Network (HAN) by employing Progressive Embedding Scale-context (PES) information, which enables the network to simultaneously suppress noise and adapt head scale variation.
10 code implementations • CVPR 2021 • Changqian Yu, Bin Xiao, Changxin Gao, Lu Yuan, Lei Zhang, Nong Sang, Jingdong Wang
We introduce a lightweight unit, conditional channel weighting, to replace costly pointwise (1x1) convolutions in shuffle blocks.
Ranked #37 on
Pose Estimation
on COCO test-dev
1 code implementation • CVPR 2021 • Xiang Wang, Shiwei Zhang, Zhiwu Qing, Yuanjie Shao, Changxin Gao, Nong Sang
In this paper, we focus on applying the power of self-supervised methods to improve semi-supervised action proposal generation.
Ranked #2 on
Semi-Supervised Action Detection
on THUMOS' 14
Self-Supervised Learning
Semi-Supervised Action Detection
+1
1 code implementation • CVPR 2021 • Zhiwu Qing, Haisheng Su, Weihao Gan, Dongliang Wang, Wei Wu, Xiang Wang, Yu Qiao, Junjie Yan, Changxin Gao, Nong Sang
In this paper, we propose Temporal Context Aggregation Network (TCANet) to generate high-quality action proposals through "local and global" temporal context aggregation and complementary as well as progressive boundary refinement.
Ranked #4 on
Temporal Action Localization
on ActivityNet-1.3
no code implementations • 22 Feb 2021 • Chuchu Han, Zhedong Zheng, Changxin Gao, Nong Sang, Yi Yang
Specifically, to reconcile the conflicts of multiple objectives, we simplify the standard tightly coupled pipelines and establish a deeply decoupled multi-task learning framework.
Ranked #7 on
Person Search
on PRW
1 code implementation • ICCV 2021 • Shizhen Zhao, Changxin Gao, Yuanjie Shao, Wei-Shi Zheng, Nong Sang
Specifically, to alleviate the intra-class variations, a clustering method is utilized to generate pseudo labels for both visual and textual instances.
1 code implementation • 17 Dec 2020 • Moran Li, Yuan Gao, Nong Sang
This is different from the previous methods where all the joints are considered holistically and share the same feature.
1 code implementation • ECCV 2020 • Shizhen Zhao, Changxin Gao, Jun Zhang, Hao Cheng, Chuchu Han, Xinyang Jiang, Xiaowei Guo, Wei-Shi Zheng, Nong Sang, Xing Sun
In the conventional person Re-ID setting, it is widely assumed that cropped person images are for each individual.
no code implementations • ECCV 2020 • Changqian Yu, Yifan Liu, Changxin Gao, Chunhua Shen, Nong Sang
In this paper, we present a Representative Graph (RepGraph) layer to dynamically sample a few representative features, which dramatically reduces redundancy.
no code implementations • 7 Aug 2020 • Xiang Wang, Changxin Gao, Shiwei Zhang, Nong Sang
By this means, the proposed MLTPN can learn rich and discriminative features for different action instances with different durations.
1 code implementation • ECCV 2020 • Yanrui Bin, Xuan Cao, Xinya Chen, Yanhao Ge, Ying Tai, Chengjie Wang, Jilin Li, Feiyue Huang, Changxin Gao, Nong Sang
Human pose estimation is the task of localizing body keypoints from still images.
no code implementations • 13 Jun 2020 • Zhiwu Qing, Xiang Wang, Yongpeng Sang, Changxin Gao, Shiwei Zhang, Nong Sang
This technical report analyzes a temporal action localization method we used in the HACS competition which is hosted in Activitynet Challenge 2020. The goal of our task is to locate the start time and end time of the action in the untrimmed video, and predict action category. Firstly, we utilize the video-level feature information to train multiple video-level action classification models.
1 code implementation • 13 Jun 2020 • Xiang Wang, Baiteng Ma, Zhiwu Qing, Yongpeng Sang, Changxin Gao, Shiwei Zhang, Nong Sang
In this report, we present our solution for the task of temporal action localization (detection) (task 1) in ActivityNet Challenge 2020.
no code implementations • 20 May 2020 • Xinya Chen, Yanrui Bin, Changxin Gao, Nong Sang, Hao Tang
The module builds a fully connected directed graph between the regions of different density where each node (region) is represented by weighted global pooled feature, and GCN is learned to map this region graph to a set of relation-aware regions representations.
1 code implementation • CVPR 2020 • Yuanjie Shao, Lerenhan Li, Wenqi Ren, Changxin Gao, Nong Sang
By training image translation and dehazing network in an end-to-end manner, we can obtain better effects of both image translation and dehazing.
Ranked #4 on
Image Dehazing
on RESIDE-6K
6 code implementations • 5 Apr 2020 • Changqian Yu, Changxin Gao, Jingbo Wang, Gang Yu, Chunhua Shen, Nong Sang
We propose to treat these spatial details and categorical semantics separately to achieve high accuracy and high efficiency for realtime semantic segmentation.
Ranked #1 on
Real-Time Semantic Segmentation
on COCO-Stuff
2 code implementations • CVPR 2020 • Changqian Yu, Jingbo Wang, Changxin Gao, Gang Yu, Chunhua Shen, Nong Sang
Given an input image and corresponding ground truth, Affinity Loss constructs an ideal affinity map to supervise the learning of Context Prior.
Ranked #1 on
Scene Understanding
on ADE20K val
1 code implementation • 19 Jan 2020 • Shizhen Zhao, Changxin Gao, Yuanjie Shao, Lerenhan Li, Changqian Yu, Zhong Ji, Nong Sang
FFU and BFU add the IoU variance to the results of CFU, yielding class-specific foreground and background features, respectively.
no code implementations • ICCV 2019 • Chuchu Han, Jiacheng Ye, Yunshan Zhong, Xin Tan, Chi Zhang, Changxin Gao, Nong Sang
The state-of-the-art methods train the detector individually, and the detected bounding boxes may be sub-optimal for the following re-ID task.
17 code implementations • ECCV 2018 • Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, Nong Sang
Semantic segmentation requires both rich spatial information and sizeable receptive field.
Ranked #4 on
Semantic Segmentation
on SkyScapes-Dense
Dichotomous Image Segmentation
Real-Time Semantic Segmentation
+1
3 code implementations • CVPR 2018 • Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, Nong Sang
Most existing methods of semantic segmentation still suffer from two aspects of challenges: intra-class inconsistency and inter-class indistinction.
Ranked #5 on
Semantic Segmentation
on PASCAL VOC 2012 test
no code implementations • CVPR 2018 • Lerenhan Li, Jinshan Pan, Wei-Sheng Lai, Changxin Gao, Nong Sang, Ming-Hsuan Yang
We present an effective blind image deblurring method based on a data-driven discriminative prior. Our work is motivated by the fact that a good image prior should favor clear images over blurred images. In this work, we formulate the image prior as a binary classifier which can be achieved by a deep convolutional neural network (CNN). The learned prior is able to distinguish whether an input image is clear or not. Embedded into the maximum a posterior (MAP) framework, it helps blind deblurring in various scenarios, including natural, face, text, and low-illumination images. However, it is difficult to optimize the deblurring method with the learned image prior as it involves a non-linear CNN. Therefore, we develop an efficient numerical approach based on the half-quadratic splitting method and gradient decent algorithm to solve the proposed model. Furthermore, the proposed model can be easily extended to non-uniform deblurring. Both qualitative and quantitative experimental results show that our method performs favorably against state-of-the-art algorithms as well as domain-specific image deblurring approaches.
no code implementations • 9 Feb 2018 • Jun Xiang, Guoshuai Zhang, Jianhua Hou, Nong Sang, Rui Huang
Designing a robust affinity model is the key issue in multiple target tracking (MTT).
no code implementations • 12 Nov 2016 • Dapeng Luo, Zhipeng Zeng, Nong Sang, Xiang Wu, Longsheng Wei, Quanzheng Mou, Jun Cheng, Chen Luo
In this paper, the proposed framework takes a remarkably different direction to resolve the multi-scene detection problem in a bottom-up fashion.
no code implementations • 29 Jun 2016 • Cong Yao, Xiang Bai, Nong Sang, Xinyu Zhou, Shuchang Zhou, Zhimin Cao
Recently, scene text detection has become an active research topic in computer vision and document analysis, because of its great importance and significant challenge.
Ranked #6 on
Scene Text Detection
on COCO-Text
no code implementations • 24 Feb 2014 • Changxin Gao, Feifei Chen, Jin-Gang Yu, Rui Huang, Nong Sang
However, the task in tracking is to search for a specific object, rather than an object category as in detection.