Search Results for author: Zehuan Yuan

Found 53 papers, 31 papers with code

Recognize Any Regions

1 code implementation2 Nov 2023 Haosen Yang, Chuofan Ma, Bin Wen, Yi Jiang, Zehuan Yuan, Xiatian Zhu

Understanding the semantics of individual regions or patches within unconstrained images, such as in open-world object detection, represents a critical yet challenging task in computer vision.

object-detection Object Recognition +1

EVE: Efficient Vision-Language Pre-training with Masked Prediction and Modality-Aware MoE

no code implementations23 Aug 2023 Junyi Chen, Longteng Guo, Jia Sun, Shuai Shao, Zehuan Yuan, Liang Lin, Dongyu Zhang

Owing to the combination of the unified architecture and pre-training task, EVE is easy to scale up, enabling better downstream performance with fewer resources and faster training speed.

Image-text matching Question Answering +5

Exploring Transformers for Open-world Instance Segmentation

no code implementations ICCV 2023 Jiannan Wu, Yi Jiang, Bin Yan, Huchuan Lu, Zehuan Yuan, Ping Luo

Open-world instance segmentation is a rising task, which aims to segment all objects in the image by learning from a limited number of base-category objects.

Contrastive Learning Instance Segmentation +1

Meta Compositional Referring Expression Segmentation

no code implementations CVPR 2023 Li Xu, Mark He Huang, Xindi Shang, Zehuan Yuan, Ying Sun, Jun Liu

Then, following a novel meta optimization scheme to optimize the model to obtain good testing performance on the virtual testing sets after training on the virtual training set, our framework can effectively drive the model to better capture semantics and visual representations of individual concepts, and thus obtain robust generalization performance even when handling novel compositions.

Meta-Learning Referring Expression +2

Token Boosting for Robust Self-Supervised Visual Transformer Pre-training

no code implementations CVPR 2023 Tianjiao Li, Lin Geng Foo, Ping Hu, Xindi Shang, Hossein Rahmani, Zehuan Yuan, Jun Liu

Pre-training VTs on such corrupted data can be challenging, especially when we pre-train via the masked autoencoding approach, where both the inputs and masked ``ground truth" targets can potentially be unreliable in this case.

Learning Instance-Level Representation for Large-Scale Multi-Modal Pretraining in E-commerce

no code implementations CVPR 2023 Yang Jin, Yongzhi Li, Zehuan Yuan, Yadong Mu

Extensive experimental results show that, without further fine-tuning, ECLIP surpasses existing methods by a large margin on a broad range of downstream tasks, demonstrating the strong transferability to real-world E-commerce applications.

Multi-Level Contrastive Learning for Dense Prediction Task

1 code implementation4 Apr 2023 Qiushan Guo, Yizhou Yu, Yi Jiang, Jiannan Wu, Zehuan Yuan, Ping Luo

We extend our pretext task to supervised pre-training, which achieves a similar performance to self-supervised learning.

Contrastive Learning Self-Supervised Learning

Universal Instance Perception as Object Discovery and Retrieval

1 code implementation CVPR 2023 Bin Yan, Yi Jiang, Jiannan Wu, Dong Wang, Ping Luo, Zehuan Yuan, Huchuan Lu

All instance perception tasks aim at finding certain objects specified by some queries such as category names, language expressions, and target annotations, but this complete field has been split into multiple independent subtasks.

Described Object Detection Generalized Referring Expression Comprehension +14

QueryPose: Sparse Multi-Person Pose Regression via Spatial-Aware Part-Level Query

3 code implementations15 Dec 2022 Yabo Xiao, Kai Su, Xiaojuan Wang, Dongdong Yu, Lei Jin, Mingshu He, Zehuan Yuan

The existing end-to-end methods rely on dense representations to preserve the spatial detail and structure for precise keypoint localization.


Self-supervised Video Representation Learning with Motion-Aware Masked Autoencoders

1 code implementation9 Oct 2022 Haosen Yang, Deng Huang, Bin Wen, Jiannan Wu, Hongxun Yao, Yi Jiang, Xiatian Zhu, Zehuan Yuan

As a result, our model can extract effectively both static appearance and dynamic motion spontaneously, leading to superior spatiotemporal representation learning capability.

Representation Learning Semantic Segmentation +2

MAMO: Masked Multimodal Modeling for Fine-Grained Vision-Language Representation Learning

no code implementations9 Oct 2022 Zijia Zhao, Longteng Guo, Xingjian He, Shuai Shao, Zehuan Yuan, Jing Liu

Our method performs joint masking on image-text input and integrates both implicit and explicit targets for the masked signals to recover.

Question Answering Representation Learning +5

Embracing Consistency: A One-Stage Approach for Spatio-Temporal Video Grounding

1 code implementation27 Sep 2022 Yang Jin, Yongzhi Li, Zehuan Yuan, Yadong Mu

Spatio-Temporal video grounding (STVG) focuses on retrieving the spatio-temporal tube of a specific object depicted by a free-form textual expression.

Spatio-Temporal Video Grounding Video Grounding

Rethinking Resolution in the Context of Efficient Video Recognition

1 code implementation26 Sep 2022 Chuofan Ma, Qiushan Guo, Yi Jiang, Zehuan Yuan, Ping Luo, Xiaojuan Qi

Our key finding is that the major cause of degradation is not information loss in the down-sampling process, but rather the mismatch between network architecture and input scale.

Knowledge Distillation Video Recognition

MCIBI++: Soft Mining Contextual Information Beyond Image for Semantic Segmentation

2 code implementations9 Sep 2022 Zhenchao Jin, Dongdong Yu, Zehuan Yuan, Lequan Yu

To this end, we propose a novel soft mining contextual information beyond image paradigm named MCIBI++ to further boost the pixel-level representations.

Segmentation Semantic Segmentation +1

Single-Stage Open-world Instance Segmentation with Cross-task Consistency Regularization

1 code implementation18 Aug 2022 Xizhe Xue, Dongdong Yu, Lingqiao Liu, Yu Liu, Satoshi Tsutsui, Ying Li, Zehuan Yuan, Ping Song, Mike Zheng Shou

Based on the single-stage instance segmentation framework, we propose a regularization model to predict foreground pixels and use its relation to instance segmentation to construct a cross-task consistency loss.

Autonomous Driving Instance Segmentation +2

You Should Look at All Objects

1 code implementation16 Jul 2022 Zhenchao Jin, Dongdong Yu, Luchuan Song, Zehuan Yuan, Lequan Yu

Feature pyramid network (FPN) is one of the key components for object detectors.

Towards Grand Unification of Object Tracking

1 code implementation14 Jul 2022 Bin Yan, Yi Jiang, Peize Sun, Dong Wang, Zehuan Yuan, Ping Luo, Huchuan Lu

We present a unified method, termed Unicorn, that can simultaneously solve four tracking problems (SOT, MOT, VOS, MOTS) with a single network using the same model parameters.

Multi-Object Tracking Multi-Object Tracking and Segmentation +2

Masked Generative Distillation

3 code implementations3 May 2022 Zhendong Yang, Zhe Li, Mingqi Shao, Dachuan Shi, Zehuan Yuan, Chun Yuan

The current distillation algorithm usually improves students' performance by imitating the output of the teacher.

Image Classification Instance Segmentation +4

Birds of A Feather Flock Together: Category-Divergence Guidance for Domain Adaptive Segmentation

no code implementations5 Apr 2022 Bo Yuan, Danpei Zhao, Shuai Shao, Zehuan Yuan, Changhu Wang

In two typical cross-domain semantic segmentation tasks, i. e., GTA5 to Cityscapes and SYNTHIA to Cityscapes, our method achieves the state-of-the-art segmentation accuracy.

Road Segmentation Segmentation +1

MetaFormer: A Unified Meta Framework for Fine-Grained Recognition

2 code implementations5 Mar 2022 Qishuai Diao, Yi Jiang, Bin Wen, Jia Sun, Zehuan Yuan

Fine-Grained Visual Classification(FGVC) is the task that requires recognizing the objects belonging to multiple subordinate categories of a super-category.

 Ranked #1 on Fine-Grained Image Classification on NABirds (using extra training data)

Fine-Grained Image Classification

Language as Queries for Referring Video Object Segmentation

1 code implementation CVPR 2022 Jiannan Wu, Yi Jiang, Peize Sun, Zehuan Yuan, Ping Luo

Referring video object segmentation (R-VOS) is an emerging cross-modal task that aims to segment the target object referred by a language expression in all video frames.

Object Tracking Referring Expression Segmentation +4

Disentangled Contrastive Learning on Graphs

no code implementations NeurIPS 2021 Haoyang Li, Xin Wang, Ziwei Zhang, Zehuan Yuan, Hang Li, Wenwu Zhu

Then we propose a novel factor-wise discrimination objective in a contrastive learning manner, which can force the factorized representations to independently reflect the expressive information from different latent factors.

Contrastive Learning Self-Supervised Learning

Trimap-guided Feature Mining and Fusion Network for Natural Image Matting

1 code implementation1 Dec 2021 Weihao Jiang, Dongdong Yu, Zhaozhi Xie, Yaoyi Li, Zehuan Yuan, Hongtao Lu

For emerging content-based feature fusion, most existing matting methods only focus on local features which lack the guidance of a global feature with strong semantic information related to the interesting object.

Image Matting

DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion

3 code implementations CVPR 2022 Peize Sun, Jinkun Cao, Yi Jiang, Zehuan Yuan, Song Bai, Kris Kitani, Ping Luo

A typical pipeline for multi-object tracking (MOT) is to use a detector for object localization, and following re-identification (re-ID) for object association.

Multi-Object Tracking object-detection +2

Focal and Global Knowledge Distillation for Detectors

1 code implementation CVPR 2022 Zhendong Yang, Zhe Li, Xiaohu Jiang, Yuan Gong, Zehuan Yuan, Danpei Zhao, Chun Yuan

Global distillation rebuilds the relation between different pixels and transfers it from teachers to students, compensating for missing global information in focal distillation.

Image Classification Knowledge Distillation +2

Multimodal Transformer with Variable-length Memory for Vision-and-Language Navigation

1 code implementation10 Nov 2021 Chuang Lin, Yi Jiang, Jianfei Cai, Lizhen Qu, Gholamreza Haffari, Zehuan Yuan

Vision-and-Language Navigation (VLN) is a task that an agent is required to follow a language instruction to navigate to the goal position, which relies on the ongoing interactions with the environment during moving.

Navigate Vision and Language Navigation

Objects in Semantic Topology

no code implementations ICLR 2022 Shuo Yang, Peize Sun, Yi Jiang, Xiaobo Xia, Ruiheng Zhang, Zehuan Yuan, Changhu Wang, Ping Luo, Min Xu

A more realistic object detection paradigm, Open-World Object Detection, has arisen increasing research interests in the community recently.

Incremental Learning Language Modelling +2

Memory Based Video Scene Parsing

no code implementations1 Sep 2021 Zhenchao Jin, Dongdong Yu, Kai Su, Zehuan Yuan, Changhu Wang

Video scene parsing is a long-standing challenging task in computer vision, aiming to assign pre-defined semantic labels to pixels of all frames in a given video.

Scene Parsing Semantic Segmentation

Center Prediction Loss for Re-identification

no code implementations30 Apr 2021 Lu Yang, Yunlong Wang, Lingqiao Liu, Peng Wang, Lu Chi, Zehuan Yuan, Changhu Wang, Yanning Zhang

In this paper, we propose a new loss based on center predictivity, that is, a sample must be positioned in a location of the feature space such that from it we can roughly predict the location of the center of same-class samples.

Conditional Hyper-Network for Blind Super-Resolution with Multiple Degradations

1 code implementation8 Apr 2021 Guanghao Yin, Wei Wang, Zehuan Yuan, Wei Ji, Dongdong Yu, Shouqian Sun, Tat-Seng Chua, Changhu Wang

We extract degradation prior at task-level with the proposed ConditionNet, which will be used to adapt the parameters of the basic SR network (BaseNet).

Blind Super-Resolution Image Super-Resolution

Exploring Balanced Feature Spaces for Representation Learning

no code implementations ICLR 2021 Bingyi Kang, Yu Li, Sa Xie, Zehuan Yuan, Jiashi Feng

Motivated by this question, we conduct a series of studies on the performance of self-supervised contrastive learning and supervised learning methods over multiple datasets where training instance distributions vary from a balanced one to a long-tailed one.

Contrastive Learning Long-tail Learning +2

Domain-Invariant Disentangled Network for Generalizable Object Detection

no code implementations ICCV 2021 Chuang Lin, Zehuan Yuan, Sicheng Zhao, Peize Sun, Changhu Wang, Jianfei Cai

By disentangling representations on both image and instance levels, DIDN is able to learn domain-invariant representations that are suitable for generalized object detection.

Disentanglement Domain Generalization +3

Unsupervised Real-World Super-Resolution: A Domain Adaptation Perspective

no code implementations ICCV 2021 Wei Wang, Haochen Zhang, Zehuan Yuan, Changhu Wang

A popular attempts towards the challenge is unpaired generative adversarial networks, which generate "real" LR counterparts from real HR images using image-to-image translation and then perform super-resolution from "real" LR->SR.

Domain Adaptation Image-to-Image Translation +1

TransTrack: Multiple Object Tracking with Transformer

2 code implementations31 Dec 2020 Peize Sun, Jinkun Cao, Yi Jiang, Rufeng Zhang, Enze Xie, Zehuan Yuan, Changhu Wang, Ping Luo

In this work, we propose TransTrack, a simple but efficient scheme to solve the multiple object tracking problems.

Ranked #4 on Multi-Object Tracking on SportsMOT (using extra training data)

Multi-Object Tracking Multiple Object Tracking with Transformer +2

Slimmable Generative Adversarial Networks

1 code implementation10 Dec 2020 Liang Hou, Zehuan Yuan, Lei Huang, HuaWei Shen, Xueqi Cheng, Changhu Wang

In particular, for real-time generation tasks, different devices require generators of different sizes due to varying computing power.

What Makes for End-to-End Object Detection?

1 code implementation10 Dec 2020 Peize Sun, Yi Jiang, Enze Xie, Wenqi Shao, Zehuan Yuan, Changhu Wang, Ping Luo

We identify that classification cost in matching cost is the main ingredient: (1) previous detectors only consider location cost, (2) by additionally introducing classification cost, previous detectors immediately produce one-to-one prediction during inference.

General Classification object-detection +1

Sparse R-CNN: End-to-End Object Detection with Learnable Proposals

6 code implementations CVPR 2021 Peize Sun, Rufeng Zhang, Yi Jiang, Tao Kong, Chenfeng Xu, Wei Zhan, Masayoshi Tomizuka, Lei LI, Zehuan Yuan, Changhu Wang, Ping Luo

In our method, however, a fixed sparse set of learned object proposals, total length of $N$, are provided to object recognition head to perform classification and location.

object-detection Object Detection +1

Controllable Orthogonalization in Training DNNs

1 code implementation CVPR 2020 Lei Huang, Li Liu, Fan Zhu, Diwen Wan, Zehuan Yuan, Bo Li, Ling Shao

Orthogonality is widely used for training deep neural networks (DNNs) due to its ability to maintain all singular values of the Jacobian close to 1 and reduce redundancy in representation.

Image Classification

Deformable Tube Network for Action Detection in Videos

no code implementations3 Jul 2019 Wei Li, Zehuan Yuan, Dashan Guo, Lei Huang, Xiangzhong Fang, Changhu Wang

To perform action detection, we design a 3D convolution network with skip connections for tube classification and regression.

Action Detection Action Recognition

Knowing Where to Look? Analysis on Attention of Visual Question Answering System

no code implementations9 Oct 2018 Wei Li, Zehuan Yuan, Xiangzhong Fang, Changhu Wang

Attention mechanisms have been widely used in Visual Question Answering (VQA) solutions due to their capacity to model deep cross-domain interactions.

Question Answering Visual Question Answering

Towards Good Practices for Multi-modal Fusion in Large-scale Video Classification

no code implementations16 Sep 2018 Jinlai Liu, Zehuan Yuan, Changhu Wang

Leveraging both visual frames and audio has been experimentally proven effective to improve large-scale video classification.

Classification General Classification +1

Temporal Action Localization by Structured Maximal Sums

no code implementations CVPR 2017 Zehuan Yuan, Jonathan C. Stroud, Tong Lu, Jia Deng

We pose action localization as a structured prediction over arbitrary-length temporal windows, where each window is scored as the sum of frame-wise classification scores.

Action Detection General Classification +2

Cannot find the paper you are looking for? You can Submit a new open access paper.