1 code implementation • 24 Oct 2024 • Haiyang Zheng, Nan Pu, Wenjing Li, Nicu Sebe, Zhun Zhong
To address these issues, we propose a novel Prototypical Hash Encoding (PHE) framework consisting of Category-aware Prototype Generation (CPG) and Discriminative Category Encoding (DCE) to mitigate the sensitivity of hash code while preserving rich discriminative information contained in high-dimension feature space, in a two-stage projection fashion.
no code implementations • 23 Oct 2024 • Yaxiong Wang, Lianwei Wu, Lechao Cheng, Zhun Zhong, Meng Wang
Recent advancements in image-text matching have been notable, yet prevailing models predominantly cater to broad queries and struggle with accommodating fine-grained query intention.
1 code implementation • 9 Oct 2024 • Shijie Ma, Fei Zhu, Zhun Zhong, Wenzhuo LIU, Xu-Yao Zhang, Cheng-Lin Liu
We delve into the conflicts and identify that models are susceptible to prediction bias and hardness bias.
no code implementations • 7 Oct 2024 • Mingxuan Liu, Zhun Zhong, Jun Li, Gianni Franchi, Subhankar Roy, Elisa Ricci
Our framework, Text Driven Semantic Multiple Clustering (TeDeSC), uses text as a proxy to concurrently reason over large image collections, discover partitioning criteria, expressed in natural language, and reveal semantic substructures.
1 code implementation • CVPR 2024 • Dong Zhao, Shuang Wang, Qi Zang, Licheng Jiao, Nicu Sebe, Zhun Zhong
Specifically, we introduce the Stable Neighbor Denoising (SND) approach, which effectively discovers highly correlated stable and unstable samples by nearest neighbor retrieval and guides the reliable optimization of unstable samples by bi-level learning.
1 code implementation • 2 Jun 2024 • Chentao Cao, Zhun Zhong, Zhanke Zhou, Yang Liu, Tongliang Liu, Bo Han
In this paper, we propose to tackle this constraint by leveraging the expert knowledge and reasoning capability of large language models (LLM) to Envision potential Outlier Exposure, termed EOE, without access to any actual OOD data.
no code implementations • 27 May 2024 • Yuting Ma, Lechao Cheng, Yaxiong Wang, Zhun Zhong, Xiaohua Xu, Meng Wang
Specifically, we employ a local prompt tuning scheme that leverages a few learnable visual prompts to efficiently fine-tune the frozen pre-trained foundation model for downstream tasks, thereby accelerating training and improving model performance under limited local resources and data heterogeneity.
1 code implementation • 26 Mar 2024 • Yuhuan Yang, Chaofan Ma, Jiangchao Yao, Zhun Zhong, Ya zhang, Yanfeng Wang
In this paper, we propose ReMamber, a novel RIS architecture that integrates the power of Mamba with a multi-modal Mamba Twister block.
1 code implementation • 20 Mar 2024 • Linshan Wu, Zhun Zhong, Jiayi Ma, Yunchao Wei, Hao Chen, Leyuan Fang, Shutao Li
Based on the label distributions, we leverage the GMM to generate high-quality pseudo labels for more reliable supervision.
Weakly supervised Semantic Segmentation Weakly-Supervised Semantic Segmentation
no code implementations • 12 Mar 2024 • Haiyang Zheng, Nan Pu, Wenjing Li, Nicu Sebe, Zhun Zhong
In this paper, we study the problem of Generalized Category Discovery (GCD), which aims to cluster unlabeled data from both known and unknown categories using the knowledge of labeled data from known categories.
2 code implementations • CVPR 2024 • Fei Wang, Dan Guo, Kun Li, Zhun Zhong, Meng Wang
To this end, we present FD4MM, a new paradigm of Frequency Decoupling for Motion Magnification with a Multi-level Isomorphic Architecture to capture multi-level high-frequency details and a stable low-frequency structure (motion field) in video space.
1 code implementation • CVPR 2024 • Shijie Ma, Fei Zhu, Zhun Zhong, Xu-Yao Zhang, Cheng-Lin Liu
Generalized Category Discovery (GCD) is a pragmatic and challenging open-world task, which endeavors to cluster unlabeled samples from both novel and old classes, leveraging some labeled data of old classes.
no code implementations • 24 Jan 2024 • Yuanpeng Tu, Zhun Zhong, Yuxi Li, Hengshuang Zhao
Generalized category discovery (GCD) aims at addressing a more realistic and challenging setting of semi-supervised learning, where only part of the category labels are assigned to certain training samples.
no code implementations • 24 Jan 2024 • Mingxuan Liu, Subhankar Roy, Wenjing Li, Zhun Zhong, Nicu Sebe, Elisa Ricci
Identifying subordinate-level categories from images is a longstanding task in computer vision and is referred to as fine-grained visual recognition (FGVR).
no code implementations • 18 Jan 2024 • Yunpeng Gong, Zhun Zhong, Yansong Qu, Zhiming Luo, Rongrong Ji, Min Jiang
For instance, infrared images are typically grayscale, unlike visible images that contain color information.
1 code implementation • 11 Dec 2023 • Dong Zhao, Ruizhi Yang, Shuang Wang, Qi Zang, Yang Hu, Licheng Jiao, Nicu Sebe, Zhun Zhong
This approach formulates pseudo-labels at the connectivity level and thus can facilitate learning structured and low-noise semantics.
no code implementations • CVPR 2024 • Nan Pu, Zhun Zhong, Xinyuan Ji, Nicu Sebe
On each client, GCL builds class-level contrastive learning with both local and global GMMs.
1 code implementation • 2 May 2023 • Juanjuan Weng, Zhiming Luo, Dazhen Lin, Shaozi Li, Zhun Zhong
Recent research has shown that Deep Neural Networks (DNNs) are highly vulnerable to adversarial samples, which are highly transferable and can be used to attack other unknown black-box models.
1 code implementation • CVPR 2023 • Nan Pu, Zhun Zhong, Nicu Sebe
This leads traditional novel category discovery (NCD) methods to be incapacitated for GCD, due to their assumption of unlabeled data are only from novel categories.
1 code implementation • 28 Mar 2023 • Mingxuan Liu, Subhankar Roy, Zhun Zhong, Nicu Sebe, Elisa Ricci
Discovering novel concepts in unlabelled datasets and in a continuous manner is an important desideratum of lifelong learners.
2 code implementations • 7 Mar 2023 • Juanjuan Weng, Zhiming Luo, Zhun Zhong, Shaozi Li, Nicu Sebe
In this work, we provide a comprehensive investigation of the CE loss function and find that the logit margin between the targeted and untargeted classes will quickly obtain saturation in CE, which largely limits the transferability.
1 code implementation • CVPR 2023 • Wei Wang, Zhun Zhong, Weijie Wang, Xi Chen, Charles Ling, Boyu Wang, Nicu Sebe
In this paper, we study the application of Test-time domain adaptation in semantic segmentation (TTDA-Seg) where both efficiency and effectiveness are crucial.
1 code implementation • CVPR 2023 • Linshan Wu, Zhun Zhong, Leyuan Fang, Xingxin He, Qiang Liu, Jiayi Ma, Hao Chen
Our AGMM can effectively endow reliable supervision for unlabeled pixels based on the distributions of labeled and unlabeled pixels.
1 code implementation • 18 Dec 2022 • Yuyang Zhao, Zhun Zhong, Na Zhao, Nicu Sebe, Gim Hee Lee
Furthermore, we present a novel style hallucination module (SHM) to generate style-diversified samples that are essential to consistency learning.
1 code implementation • 18 Jul 2022 • Subhankar Roy, Mingxuan Liu, Zhun Zhong, Nicu Sebe, Elisa Ricci
We study the new task of class-incremental Novel Class Discovery (class-iNCD), which refers to the problem of discovering novel categories in an unlabelled data set by leveraging a pre-trained model that has been trained on a labelled data set containing disjoint yet related categories.
1 code implementation • 11 Jul 2022 • Zhun Zhong, Yuyang Zhao, Gim Hee Lee, Nicu Sebe
Experiments on two synthetic-to-real semantic segmentation benchmarks demonstrate that AdvStyle can significantly improve the model performance on unseen real domains and show that we can achieve the state of the art.
1 code implementation • IEEE Transactions on Image Processing (TIP) 2022 • Jinliang Lin, Zhedong Zheng, Zhun Zhong, Zhiming Luo, Shaozi Li, Yi Yang, Nicu Sebe
Inspired by the human visual system for mining local patterns, we propose a new framework called RK-Net to jointly learn the discriminative Representation and detect salient Keypoints with a single Network.
Ranked #2 on Drone navigation on University-1652
2 code implementations • 6 Apr 2022 • Yuyang Zhao, Zhun Zhong, Na Zhao, Nicu Sebe, Gim Hee Lee
Furthermore, we present a novel style hallucination module (SHM) to generate style-diversified samples that are essential to consistency learning.
Ranked #4 on Robust Object Detection on DWD
no code implementations • 5 Mar 2022 • Fengxiang Yang, Zhun Zhong, Zhiming Luo, Shaozi Li, Nicu Sebe
During local training, the DFS are used to synthesize novel domain statistics with the proposed domain hallucinating, which is achieved by re-weighting DFS with random weights.
no code implementations • 3 Mar 2022 • Yongguo Ling, Zhun Zhong, Donglin Cao, Zhiming Luo, Yaojin Lin, Shaozi Li, Nicu Sebe
In this manner, the model will focus on reducing the inter-modality discrepancy while paying less attention to intra-identity variations, leading to a more effective modality alignment.
1 code implementation • CVPR 2022 • Yuyang Zhao, Zhun Zhong, Nicu Sebe, Gim Hee Lee
We introduce a new setting of Novel Class Discovery in Semantic Segmentation (NCDSS), which aims at segmenting unlabeled images containing new classes given prior knowledge from a labeled set of disjoint classes.
1 code implementation • 2 Dec 2021 • Jichao Zhang, Enver Sangineto, Hao Tang, Aliaksandr Siarohin, Zhun Zhong, Nicu Sebe, Wei Wang
However, they usually struggle to generate high-quality images representing non-rigid objects, such as the human body, which is of a great interest for many computer graphics applications.
1 code implementation • 19 Nov 2021 • Guanglei Yang, Zhun Zhong, Hao Tang, Mingli Ding, Nicu Sebe, Elisa Ricci
Specifically, in the image translation stage, Bi-Mix leverages the knowledge of day-night image pairs to improve the quality of nighttime image relighting.
1 code implementation • Neurocomputing 2021 • Nanyu Li, Junzhen Wang, Zhimin Luo, Zhun Zhong, Shaozi Li
Of those, methods based on bilinear pooling are one of the main categories for computing the interaction between deep features and have shown high effectiveness.
Ranked #1 on Fine-Grained Image Recognition on CUB Birds
Fine-Grained Image Classification Fine-Grained Image Recognition +3
1 code implementation • ICCV 2021 • Enrico Fini, Enver Sangineto, Stéphane Lathuilière, Zhun Zhong, Moin Nabi, Elisa Ricci
In this paper, we study the problem of Novel Class Discovery (NCD).
Ranked #3 on Novel Object Detection on LVIS v1.0 val
1 code implementation • CVPR 2021 • Zhun Zhong, Enrico Fini, Subhankar Roy, Zhiming Luo, Elisa Ricci, Nicu Sebe
In this paper, we address Novel Class Discovery (NCD), the task of unveiling new classes in a set of unlabeled samples given a labeled dataset with known classes.
1 code implementation • 7 Jun 2021 • Yuyang Zhao, Zhun Zhong, Zhiming Luo, Gim Hee Lee, Nicu Sebe
Second, CPSS can reduce the influence of noisy pseudo-labels and also avoid the model overfitting to the target domain during self-supervised learning, consistently boosting the performance on the target and open domains.
1 code implementation • 28 May 2021 • Guanglei Yang, Hao Tang, Zhun Zhong, Mingli Ding, Ling Shao, Nicu Sebe, Elisa Ricci
In this paper, we study the task of source-free domain adaptation (SFDA), where the source data are not available during target adaptation.
1 code implementation • CVPR 2021 • Subhankar Roy, Evgeny Krivosheev, Zhun Zhong, Nicu Sebe, Elisa Ricci
In this paper we address multi-target domain adaptation (MTDA), where given one labeled source dataset and multiple unlabeled target datasets that differ in data distributions, the task is to learn a robust predictor for all the target domains.
Ranked #2 on Multi-target Domain Adaptation on Office-Home
Blended-target Domain Adaptation Multi-target Domain Adaptation
1 code implementation • CVPR 2021 • Fengxiang Yang, Zhun Zhong, Zhiming Luo, Yuanzheng Cai, Yaojin Lin, Shaozi Li, Nicu Sebe
This paper considers the problem of unsupervised person re-identification (re-ID), which aims to learn discriminative models with unlabeled data.
1 code implementation • CVPR 2021 • Yuyang Zhao, Zhun Zhong, Fengxiang Yang, Zhiming Luo, Yaojin Lin, Shaozi Li, Nicu Sebe
In this paper, we study the problem of multi-source domain generalization in ReID, which aims to learn a model that can perform well on unseen domains with only several labeled source domains.
no code implementations • CVPR 2021 • Zhun Zhong, Linchao Zhu, Zhiming Luo, Shaozi Li, Yi Yang, Nicu Sebe
In this paper, we tackle the problem of discovering new classes in unlabeled visual data given labeled data from disjoint classes.
1 code implementation • 3 Dec 2019 • Fengxiang Yang, Ke Li, Zhun Zhong, Zhiming Luo, Xing Sun, Hao Cheng, Xiaowei Guo, Feiyue Huang, Rongrong Ji, Shaozi Li
This procedure encourages that the selected training samples can be both clean and miscellaneous, and that the two models can promote each other iteratively.
Ranked #10 on Unsupervised Domain Adaptation on Market to Duke
no code implementations • 1 Aug 2019 • Zhun Zhong, Liang Zheng, Zhiming Luo, Shaozi Li, Yi Yang
This work considers the problem of unsupervised domain adaptation in person re-identification (re-ID), which aims to transfer knowledge from the source domain to the target domain.
Ranked #8 on Unsupervised Domain Adaptation on Market to MSMT
2 code implementations • CVPR 2019 • Zhun Zhong, Liang Zheng, Zhiming Luo, Shaozi Li, Yi Yang
To achieve this goal, an exemplar memory is introduced to store features of the target domain and accommodate the three invariance properties.
Domain Adaptive Person Re-Identification Person Re-Identification +1
1 code implementation • 5 Nov 2018 • Fengxiang Yang, Zhun Zhong, Zhiming Luo, Sheng Lian, Shaozi Li
For training of deep re-ID model, we divide it into three steps: 1) pre-training a coarse re-ID model by using virtual data; 2) collaborative filtering based positive pair mining from the real data; and 3) fine-tuning of the coarse re-ID model by leveraging the mined positive pairs and virtual data.
1 code implementation • ECCV 2018 • Zhun Zhong, Liang Zheng, Shaozi Li, Yi Yang
Person re-identification (re-ID) poses unique challenges for unsupervised domain adaptation (UDA) in that classes in the source and target sets (domains) are entirely different and that image variations are largely caused by cameras.
1 code implementation • 30 Jan 2018 • Qingji Guan, Yaping Huang, Zhun Zhong, Zhedong Zheng, Liang Zheng, Yi Yang
This paper considers the task of thorax disease classification on chest X-ray images.
10 code implementations • CVPR 2018 • Zhun Zhong, Liang Zheng, Zhedong Zheng, Shaozi Li, Yi Yang
In this paper, we explicitly consider this challenge by introducing camera style (CamStyle) adaptation.
Ranked #72 on Person Re-Identification on DukeMTMC-reID
18 code implementations • 16 Aug 2017 • Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, Yi Yang
In this paper, we introduce Random Erasing, a new data augmentation method for training the convolutional neural network (CNN).
Ranked #3 on Image Classification on Fashion-MNIST
no code implementations • CVPR 2017 • Zhun Zhong, Liang Zheng, Donglin Cao, Shaozi Li
Specifically, given an image, a k-reciprocal feature is calculated by encoding its k-reciprocal nearest neighbors into a single vector, which is used for re-ranking under the Jaccard distance.
Ranked #13 on Person Re-Identification on CUHK03
no code implementations • 19 May 2016 • Zhun Zhong, Mingyi Lei, Shaozi Li, Jianping Fan
In this paper, we propose a semantic, class-specific approach to re-rank object proposals, which can consistently improve the recall performance even with less proposals.
no code implementations • 8 May 2016 • Zhun Zhong, Songzhi Su, Donglin Cao, Shaozi Li
Secondly, we present a ground control points selection scheme according to the maximum matching confidence of each pixel.
no code implementations • 3 Jun 2015 • Zhun Zhong, Zongmin Li, Runlin Li, Xiaoxia Sun
However, when the data instances of a target domain have a different distribution than that of a source domain, the dictionary learning method may fail to perform well.