Search Results for author: Zhaoxiang Zhang

Found 76 papers, 45 papers with code

Driving into the Future: Multiview Visual Forecasting and Planning with World Model for Autonomous Driving

1 code implementation29 Nov 2023 Yuqi Wang, JiaWei He, Lue Fan, Hongxin Li, Yuntao Chen, Zhaoxiang Zhang

In autonomous driving, predicting future events in advance and evaluating the foreseeable risks empowers autonomous vehicles to better plan their actions, enhancing safety and efficiency on the road.

Autonomous Driving

Expanding Scene Graph Boundaries: Fully Open-vocabulary Scene Graph Generation via Visual-Concept Alignment and Retention

no code implementations18 Nov 2023 Zuyao Chen, Jinlin Wu, Zhen Lei, Zhaoxiang Zhang, Changwen Chen

For the more challenging settings of relation-involved open vocabulary SGG, the proposed approach integrates relation-aware pre-training utilizing image-caption data and retains visual-concept alignment through knowledge distillation.

Concept Alignment Graph Generation +4

Visual Commonsense based Heterogeneous Graph Contrastive Learning

no code implementations11 Nov 2023 Zongzhao Li, Xiangyu Zhu, Xi Zhang, Zhaoxiang Zhang, Zhen Lei

Specifically, our model contains two key components: the Commonsense-based Contrastive Learning and the Graph Relation Network.

Contrastive Learning Question Answering +3

RoleLLM: Benchmarking, Eliciting, and Enhancing Role-Playing Abilities of Large Language Models

1 code implementation1 Oct 2023 Zekun Moore Wang, Zhongyuan Peng, Haoran Que, Jiaheng Liu, Wangchunshu Zhou, Yuhan Wu, Hongcheng Guo, Ruitong Gan, Zehao Ni, Man Zhang, Zhaoxiang Zhang, Wanli Ouyang, Ke Xu, Wenhu Chen, Jie Fu, Junran Peng

The advent of Large Language Models (LLMs) has paved the way for complex tasks such as role-playing, which enhances user interactions by enabling models to imitate various characters.


Informative Data Mining for One-Shot Cross-Domain Semantic Segmentation

no code implementations ICCV 2023 Yuxi Wang, Jian Liang, Jun Xiao, Shuqi Mei, Yuran Yang, Zhaoxiang Zhang

One-shot domain adaptation methods attempt to overcome these challenges by transferring the pre-trained source model to the target domain using only one target data.

Domain Adaptation Semantic Segmentation +1

DropPos: Pre-Training Vision Transformers by Reconstructing Dropped Positions

1 code implementation NeurIPS 2023 Haochen Wang, Junsong Fan, Yuxi Wang, Kaiyou Song, Tong Wang, Zhaoxiang Zhang

As it is empirically observed that Vision Transformers (ViTs) are quite insensitive to the order of input tokens, the need for an appropriate self-supervised pretext task that enhances the location awareness of ViTs is becoming evident.

AnoVL: Adapting Vision-Language Models for Unified Zero-shot Anomaly Localization

1 code implementation30 Aug 2023 Hanqiu Deng, Zhaoxiang Zhang, Jinan Bao, Xingyu Li

On top of the proposed TFA, we further introduce a test-time adaptation (TTA) mechanism to refine anomaly localization results, where a layer of trainable parameters in the adapter is optimized using TFA's pseudo-labels and synthetic noise-corrupted tokens.

Anomaly Detection zero-shot anomaly detection

FSD V2: Improving Fully Sparse 3D Object Detection with Virtual Voxels

2 code implementations7 Aug 2023 Lue Fan, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

Consequently, we develop a suite of components to complement the virtual voxel concept, including a virtual voxel encoder, a virtual voxel mixer, and a virtual voxel assignment strategy.

3D Object Detection Clustering +4

DiffusePast: Diffusion-based Generative Replay for Class Incremental Semantic Segmentation

no code implementations2 Aug 2023 Jingfan Chen, Yuxi Wang, Pengfei Wang, Xiao Chen, Zhaoxiang Zhang, Zhen Lei, Qing Li

The Class Incremental Semantic Segmentation (CISS) extends the traditional segmentation task by incrementally learning newly added classes.

Class-Incremental Semantic Segmentation Segmentation

DDG-Net: Discriminability-Driven Graph Network for Weakly-supervised Temporal Action Localization

1 code implementation ICCV 2023 Xiaojun Tang, Junsong Fan, Chuanchen Luo, Zhaoxiang Zhang, Man Zhang, Zongyuan Yang

Considering this phenomenon, we propose Discriminability-Driven Graph Network (DDG-Net), which explicitly models ambiguous snippets and discriminative snippets with well-designed connections, preventing the transmission of ambiguous information and enhancing the discriminability of snippet-level representations.

Weakly-supervised Temporal Action Localization Weakly Supervised Temporal Action Localization

BMAD: Benchmarks for Medical Anomaly Detection

1 code implementation20 Jun 2023 Jinan Bao, Hanshi Sun, Hanqiu Deng, Yinsheng He, Zhaoxiang Zhang, Xingyu Li

However, there is a lack of a universal and fair benchmark for evaluating AD methods on medical images, which hinders the development of more generalized and robust AD methods in this specific domain.

Anomaly Detection Medical Diagnosis

Visually-Guided Sound Source Separation with Audio-Visual Predictive Coding

1 code implementation19 Jun 2023 Zengjie Song, Zhaoxiang Zhang

The framework of visually-guided sound source separation generally consists of three parts: visual feature extraction, multimodal feature fusion, and sound signal processing.

valid Visually Guided Sound Source Separation

PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic Segmentation

1 code implementation16 Jun 2023 Yuqi Wang, Yuntao Chen, Xingyu Liao, Lue Fan, Zhaoxiang Zhang

In this work, we address this limitation by studying camera-based 3D panoptic segmentation, aiming to achieve a unified occupancy representation for camera-only 3D scene understanding.

Autonomous Driving object-detection +5

Tracking Objects with 3D Representation from Videos

no code implementations8 Jun 2023 JiaWei He, Lue Fan, Yuqi Wang, Yuntao Chen, Zehao Huang, Naiyan Wang, Zhaoxiang Zhang

In this paper, we rethink the data association in 2D MOT and utilize the 3D object representation to separate each object in the feature space.

Multiple Object Tracking Representation Learning

Using Unreliable Pseudo-Labels for Label-Efficient Semantic Segmentation

no code implementations4 Jun 2023 Haochen Wang, Yuchao Wang, Yujun Shen, Junsong Fan, Yuxi Wang, Zhaoxiang Zhang

A common practice is to select the highly confident predictions as the pseudo-ground-truths for each pixel, but it leads to a problem that most pixels may be left unused due to their unreliability.

Semantic Segmentation

Pulling Target to Source: A New Perspective on Domain Adaptive Semantic Segmentation

no code implementations23 May 2023 Haochen Wang, Yujun Shen, Jingjing Fei, Wei Li, Liwei Wu, Yuxi Wang, Zhaoxiang Zhang

To this end, we propose T2S-DA, which we interpret as a form of pulling Target to Source for Domain Adaptation, encouraging the model in learning similar cross-domain features.

Domain Generalization Semantic Segmentation

Fully Sparse Fusion for 3D Object Detection

1 code implementation24 Apr 2023 Yingyan Li, Lue Fan, Yang Liu, Zehao Huang, Yuntao Chen, Naiyan Wang, Zhaoxiang Zhang, Tieniu Tan

In this paper, we study how to effectively leverage image modality in the emerging fully sparse architecture.

3D Instance Segmentation 3D Object Detection +2

Once Detected, Never Lost: Surpassing Human Performance in Offline LiDAR based 3D Object Detection

2 code implementations ICCV 2023 Lue Fan, Yuxue Yang, Yiming Mao, Feng Wang, Yuntao Chen, Naiyan Wang, Zhaoxiang Zhang

Drawing inspiration from this, we propose a high-performance offline detector in a track-centric perspective instead of the conventional object-centric perspective.

3D Object Detection object-detection

Hard Patches Mining for Masked Image Modeling

1 code implementation CVPR 2023 Haochen Wang, Kaiyou Song, Junsong Fan, Yuxi Wang, Jin Xie, Zhaoxiang Zhang

We observe that the reconstruction loss can naturally be the metric of the difficulty of the pre-training task.

Learnable Graph Matching: A Practical Paradigm for Data Association

1 code implementation27 Mar 2023 JiaWei He, Zehao Huang, Naiyan Wang, Zhaoxiang Zhang

For image matching, our method outperforms state-of-the-art methods with half training data and iterations on a popular indoor dataset, ScanNet.

Graph Matching Multiple Object Tracking +1

Sharpness-Aware Gradient Matching for Domain Generalization

1 code implementation CVPR 2023 Pengfei Wang, Zhaoxiang Zhang, Zhen Lei, Lei Zhang

In this paper, we present two conditions to ensure that the model could converge to a flat minimum with a small loss, and present an algorithm, named Sharpness-Aware Gradient Matching (SAGM), to meet the two conditions for improving model generalization capability.

Domain Generalization

A Survey of Deep Visual Cross-Domain Few-Shot Learning

no code implementations16 Mar 2023 Wenjian Wang, Lijuan Duan, Yuxi Wang, Junsong Fan, Zhi Gong, Zhaoxiang Zhang

Research into Cross-Domain Few-Shot (CDFS) has emerged to address this issue, forming a more challenging and realistic setting.

cross-domain few-shot learning Transfer Learning

Blind Video Deflickering by Neural Filtering with a Flawed Atlas

1 code implementation CVPR 2023 Chenyang Lei, Xuanchi Ren, Zhaoxiang Zhang, Qifeng Chen

Prior work usually requires specific guidance such as the flickering frequency, manual annotations, or extra consistent videos to remove the flicker.

Video Generation Video Temporal Consistency

Intrinsic Physical Concepts Discovery with Object-Centric Predictive Models

no code implementations CVPR 2023 Qu Tang, Xiangyu Zhu, Zhen Lei, Zhaoxiang Zhang

The ability to discover abstract physical concepts and understand how they work in the world through observing lies at the core of human intelligence.

Fairly Adaptive Negative Sampling for Recommendations

no code implementations16 Feb 2023 Xiao Chen, Wenqi Fan, Jingfan Chen, Haochen Liu, Zitao Liu, Zhaoxiang Zhang, Qing Li

Pairwise learning strategies are prevalent for optimizing recommendation models on implicit feedback data, which usually learns user preference by discriminating between positive (i. e., clicked by a user) and negative items (i. e., obtained by negative sampling).


Super Sparse 3D Object Detection

2 code implementations5 Jan 2023 Lue Fan, Yuxue Yang, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

To enable efficient long-range detection, we first propose a fully sparse object detector termed FSD.

3D Object Detection Autonomous Driving +1

Extracting Semantic Knowledge from GANs with Unsupervised Learning

no code implementations30 Nov 2022 Jianjin Xu, Zhaoxiang Zhang, Xiaolin Hu

Second, we train image-to-image translation networks on the synthesized datasets, enabling semantic-conditional image synthesis without human annotations.

Image Segmentation Image-to-Image Translation +2

RRSR:Reciprocal Reference-based Image Super-Resolution with Progressive Feature Alignment and Selection

no code implementations8 Nov 2022 Lin Zhang, Xin Li, Dongliang He, Fu Li, Yili Wang, Zhaoxiang Zhang

While previous state-of-the-art RefSR methods mainly focus on improving the efficacy and robustness of reference feature transfer, it is generally overlooked that a well reconstructed SR image should enable better SR reconstruction for its similar LR images when it is referred to as.

feature selection Image Super-Resolution

Pointly-Supervised Panoptic Segmentation

1 code implementation25 Oct 2022 Junsong Fan, Zhaoxiang Zhang, Tieniu Tan

In this paper, we propose a new approach to applying point-level annotations for weakly-supervised panoptic segmentation.

Panoptic Segmentation Segmentation +3

4D Unsupervised Object Discovery

1 code implementation10 Oct 2022 Yuqi Wang, Yuntao Chen, Zhaoxiang Zhang

In this paper, we propose 4D unsupervised object discovery, jointly discovering objects from 4D data -- 3D point clouds and 2D RGB images with temporal information.

3D Instance Segmentation object-detection +3

MemoNav: Selecting Informative Memories for Visual Navigation

no code implementations20 Aug 2022 Hongxin Li, Xu Yang, Yuran Yang, Shuqi Mei, Zhaoxiang Zhang

To address this limitation, we present the MemoNav, a novel memory mechanism for image-goal navigation, which retains the agent's informative short-term memory and long-term memory to improve the navigation performance on a multi-goal task.

Action Generation Graph Attention +2

Pro-tuning: Unified Prompt Tuning for Vision Tasks

no code implementations28 Jul 2022 Xing Nie, Bolin Ni, Jianlong Chang, Gaomeng Meng, Chunlei Huo, Zhaoxiang Zhang, Shiming Xiang, Qi Tian, Chunhong Pan

To this end, we propose parameter-efficient Prompt tuning (Pro-tuning) to adapt frozen vision models to various downstream vision tasks.

Adversarial Robustness Image Classification +4

Fully Sparse 3D Object Detection

4 code implementations20 Jul 2022 Lue Fan, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

To enable efficient long-range LiDAR-based object detection, we build a fully sparse 3D object detector (FSD).

3D Object Detection Autonomous Driving

Densely Constrained Depth Estimator for Monocular 3D Object Detection

1 code implementation20 Jul 2022 Yingyan Li, Yuntao Chen, JiaWei He, Zhaoxiang Zhang

So these methods only use a small number of projection constraints and produce insufficient depth candidates, leading to inaccurate depth estimation.

Depth Estimation Graph Matching +2

Implicit Sample Extension for Unsupervised Person Re-Identification

1 code implementation CVPR 2022 Xinyu Zhang, Dongdong Li, Zhigang Wang, Jian Wang, Errui Ding, Javen Qinfeng Shi, Zhaoxiang Zhang, Jingdong Wang

Specifically, we generate support samples from actual samples and their neighbouring clusters in the embedding space through a progressive linear interpolation (PLI) strategy.

Clustering Unsupervised Person Re-Identification

HP-Capsule: Unsupervised Face Part Discovery by Hierarchical Parsing Capsule Network

no code implementations CVPR 2022 Chang Yu, Xiangyu Zhu, Xiaomei Zhang, Zidu Wang, Zhaoxiang Zhang, Zhen Lei

Capsule networks are designed to present the objects by a set of parts and their relationships, which provide an insight into the procedure of visual perception.

DATA: Domain-Aware and Task-Aware Self-supervised Learning

1 code implementation CVPR 2022 Qing Chang, Junran Peng, Lingxie Xie, Jiajun Sun, Haoran Yin, Qi Tian, Zhaoxiang Zhang

However, due to the high training costs and the unconsciousness of downstream usages, most self-supervised learning methods lack the capability to correspond to the diversities of downstream scenarios, as there are various data domains, different vision tasks and latency constraints on models.

Image Classification Model Selection +5

The Devil Is in the Details: Window-based Attention for Image Compression

2 code implementations CVPR 2022 Renjie Zou, Chunfeng Song, Zhaoxiang Zhang

Inspired by recent progresses of Vision Transformer (ViT) and Swin Transformer, we found that combining the local-aware attention mechanism with the global-related feature learning could meet the expectation in image compression.

Image Compression

Emergence of Machine Language: Towards Symbolic Intelligence with Neural Networks

no code implementations14 Jan 2022 Yuqi Wang, Xu-Yao Zhang, Cheng-Lin Liu, Zhaoxiang Zhang

Moreover, through experiments we show that discrete language representation has several advantages compared with continuous feature representation, from the aspects of interpretability, generalization, and robustness.

Remember the Difference: Cross-Domain Few-Shot Semantic Segmentation via Meta-Memory Transfer

no code implementations CVPR 2022 Wenjian Wang, Lijuan Duan, Yuxi Wang, Qing En, Junsong Fan, Zhaoxiang Zhang

To remedy this problem, we propose an interesting and challenging cross-domain few-shot semantic segmentation task, where the training and test tasks perform on different domains.

Contrastive Learning Cross-Domain Few-Shot +3

Continual Stereo Matching of Continuous Driving Scenes With Growing Architecture

1 code implementation CVPR 2022 Chenghao Zhang, Kun Tian, Bin Fan, Gaofeng Meng, Zhaoxiang Zhang, Chunhong Pan

The deep stereo models have achieved state-of-the-art performance on driving scenes, but they suffer from severe performance degradation when tested on unseen scenes.

Continual Learning Stereo Matching

Embracing Single Stride 3D Object Detector with Sparse Transformer

2 code implementations CVPR 2022 Lue Fan, Ziqi Pang, Tianyuan Zhang, Yu-Xiong Wang, Hang Zhao, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

In LiDAR-based 3D object detection for autonomous driving, the ratio of the object size to input scene size is significantly smaller compared to 2D detection cases.

3D Object Detection Autonomous Driving +2

Immortal Tracker: Tracklet Never Dies

1 code implementation26 Nov 2021 Qitai Wang, Yuntao Chen, Ziqi Pang, Naiyan Wang, Zhaoxiang Zhang

We employ a simple Kalman filter for trajectory prediction and preserve the tracklet by prediction when the target is not visible.

3D Multi-Object Tracking Trajectory Prediction


no code implementations ICLR 2022 Qu Tang, Xiangyu Zhu, Zhen Lei, Zhaoxiang Zhang

In this paper, we work on object dynamics and propose Object Dynamics Distillation Network (ODDN), a framework that distillates explicit object dynamics (e. g., velocity) from sequential static representations.

Predict Future Video Frames Video Understanding

GAIA: A Transfer Learning System of Object Detection that Fits Your Needs

1 code implementation CVPR 2021 Xingyuan Bu, Junran Peng, Junjie Yan, Tieniu Tan, Zhaoxiang Zhang

Transfer learning with pre-training on large-scale datasets has played an increasingly significant role in computer vision and natural language processing recently.

object-detection Object Detection +1

Image Inpainting by End-to-End Cascaded Refinement with Mask Awareness

1 code implementation28 Apr 2021 Manyu Zhu, Dongliang He, Xin Li, Chao Li, Fu Li, Xiao Liu, Errui Ding, Zhaoxiang Zhang

Inpainting arbitrary missing regions is challenging because learning valid features for various masked regions is nontrivial.

Image Inpainting valid

Distractor-Aware Fast Tracking via Dynamic Convolutions and MOT Philosophy

1 code implementation CVPR 2021 Zikai Zhang, Bineng Zhong, Shengping Zhang, Zhenjun Tang, Xin Liu, Zhaoxiang Zhang

A practical long-term tracker typically contains three key properties, i. e. an efficient model design, an effective global re-detection strategy and a robust distractor awareness mechanism.

Multiple Object Tracking Philosophy

RefineMask: Towards High-Quality Instance Segmentation with Fine-Grained Features

1 code implementation CVPR 2021 Gang Zhang, Xin Lu, Jingru Tan, Jianmin Li, Zhaoxiang Zhang, Quanquan Li, Xiaolin Hu

In this work, we propose a new method called RefineMask for high-quality instance segmentation of objects and scenes, which incorporates fine-grained features during the instance-wise segmenting process in a multi-stage manner.

Instance Segmentation Semantic Segmentation +1

Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression

2 code implementations CVPR 2021 Zigang Geng, Ke Sun, Bin Xiao, Zhaoxiang Zhang, Jingdong Wang

Our motivation is that regressing keypoint positions accurately needs to learn representations that focus on the keypoint regions.

Keypoint Detection regression

RangeDet:In Defense of Range View for LiDAR-based 3D Object Detection

1 code implementation18 Mar 2021 Lue Fan, Xuan Xiong, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

The most notable difference with previous works is that our method is purely based on the range view representation.

3D Object Detection object-detection +2

RangeDet: In Defense of Range View for LiDAR-Based 3D Object Detection

1 code implementation ICCV 2021 Lue Fan, Xuan Xiong, Feng Wang, Naiyan Wang, Zhaoxiang Zhang

We first analyze the existing range-view-based methods and find two issues overlooked by previous works: 1) the scale variation between nearby and far away objects; 2) the inconsistency between the 2D range image coordinates used in feature extraction and the 3D Cartesian coordinates used in output.

3D Object Detection object-detection +2

Uncertainty-Aware Pseudo Label Refinery for Domain Adaptive Semantic Segmentation

no code implementations ICCV 2021 Yuxi Wang, Junran Peng, Zhaoxiang Zhang

Unsupervised domain adaptation for semantic segmentation aims to assign the pixel-level labels for unlabeled target domain by transferring knowledge from the labeled source domain.

Pseudo Label Self-Supervised Learning +2

Clothing Status Awareness for Long-Term Person Re-Identification

no code implementations ICCV 2021 Yan Huang, Qiang Wu, Jingsong Xu, Yi Zhong, Zhaoxiang Zhang

This work argues that these approaches in fact are not aware of clothing status (i. e., change or no-change) of a pedestrian.

Person Re-Identification

Group-Wise Semantic Mining for Weakly Supervised Semantic Segmentation

1 code implementation9 Dec 2020 Xueyi Li, Tianfei Zhou, Jianwu Li, Yi Zhou, Zhaoxiang Zhang

We formulate WSSS as a novel group-wise learning task that explicitly models semantic dependencies in a group of images to estimate more reliable pseudo ground-truths, which can be used for training more accurate segmentation models.

Ranked #29 on Weakly-Supervised Semantic Segmentation on COCO 2014 val (using extra training data)

Segmentation Structured Prediction +2

Unsupervised Object Detection with LiDAR Clues

no code implementations CVPR 2021 Hao Tian, Yuntao Chen, Jifeng Dai, Zhaoxiang Zhang, Xizhou Zhu

We further identify another major issue, seldom noticed by the community, that the long-tailed and open-ended (sub-)category distribution should be accommodated.

object-detection Object Detection +1

Manual-Label Free 3D Detection via An Open-Source Simulator

no code implementations16 Nov 2020 Zhen Yang, Chi Zhang, Huiming Guo, Zhaoxiang Zhang

In this paper, we propose a manual-label free 3D detection algorithm that leverages the CARLA simulator to generate a large amount of self-labeled training samples and introduces a novel Domain Adaptive VoxelNet (DA-VoxelNet) that can cross the distribution gap from the synthetic data to the real scenario.

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