915 citations · 3.2k across the 107 of their papers we have counts for
108 papers
Pipeline MoE: A Flexible MoE Implementation with Pipeline Parallelism
Xin Chen, Hengheng Zhang, Xiaotao Gu +3
The Mixture of Experts (MoE) model becomes an important choice of large language models nowadays because of its scalability with sublinear computational complexity for training and…
Segment Anything in 3D with Radiance Fields
Jiazhong Cen, Jiemin Fang, Zanwei Zhou +5
The Segment Anything Model (SAM) emerges as a powerful vision foundation model to generate high-quality 2D segmentation results. This paper aims to generalize SAM to segment 3D obj…
Focus on Your Target: A Dual Teacher-Student Framework for Domain-adaptive Semantic Segmentation
Xinyue Huo, Lingxi Xie, Wengang Zhou +2
We study unsupervised domain adaptation (UDA) for semantic segmentation. Currently, a popular UDA framework lies in self-training which endows the model with two-fold abilities: (i…
USAGE: A Unified Seed Area Generation Paradigm for Weakly Supervised Semantic Segmentation
Zelin Peng, Guanchun Wang, Lingxi Xie +3
Seed area generation is usually the starting point of weakly supervised semantic segmentation (WSSS). Computing the Class Activation Map (CAM) from a multi-label classification net…
Feature Calibration Network for Occluded Pedestrian Detection
Tianliang Zhang, Qixiang Ye, Baochang Zhang +3
Pedestrian detection in the wild remains a challenging problem especially for scenes containing serious occlusion. In this paper, we propose a novel feature learning method in the…
Pangu-Weather: A 3D High-Resolution Model for Fast and Accurate Global Weather Forecast
Kaifeng Bi, Lingxi Xie, Hengheng Zhang +3
In this paper, we present Pangu-Weather, a deep learning based system for fast and accurate global weather forecast. For this purpose, we establish a data-driven environment by dow…
Learnable Distribution Calibration for Few-Shot Class-Incremental Learning
Binghao Liu, Boyu Yang, Lingxi Xie +3
Few-shot class-incremental learning (FSCIL) faces challenges of memorizing old class distributions and estimating new class distributions given few training samples. In this study,…
Skeleton-Parted Graph Scattering Networks for 3D Human Motion Prediction
Maosen Li, Siheng Chen, Zijing Zhang +3
Graph convolutional network based methods that model the body-joints' relations, have recently shown great promise in 3D skeleton-based human motion prediction. However, these meth…
Active Pointly-Supervised Instance Segmentation
Chufeng Tang, Lingxi Xie, Gang Zhang +3
The requirement of expensive annotations is a major burden for training a well-performed instance segmentation model. In this paper, we present an economic active learning setting,…
Visual Recognition by Request
Chufeng Tang, Lingxi Xie, Xiaopeng Zhang +2
Humans have the ability of recognizing visual semantics in an unlimited granularity, but existing visual recognition algorithms cannot achieve this goal. In this paper, we establis…
A Survey on Label-efficient Deep Image Segmentation: Bridging the Gap between Weak Supervision and Dense Prediction
Wei Shen, Zelin Peng, Xuehui Wang +6
The rapid development of deep learning has made a great progress in image segmentation, one of the fundamental tasks of computer vision. However, the current segmentation algorithm…
HiViT: Hierarchical Vision Transformer Meets Masked Image Modeling
Xiaosong Zhang, Yunjie Tian, Wei Huang +4
Recently, masked image modeling (MIM) has offered a new methodology of self-supervised pre-training of vision transformers. A key idea of efficient implementation is to discard the…
Fast Dynamic Radiance Fields with Time-Aware Neural Voxels
Jiemin Fang, Taoran Yi, Xinggang Wang +5
Neural radiance fields (NeRF) have shown great success in modeling 3D scenes and synthesizing novel-view images. However, most previous NeRF methods take much time to optimize one…
Domain-Agnostic Prior for Transfer Semantic Segmentation
Xinyue Huo, Lingxi Xie, Hengtong Hu +3
Unsupervised domain adaptation (UDA) is an important topic in the computer vision community. The key difficulty lies in defining a common property between the source and target dom…
CenterNet++ for Object Detection
Kaiwen Duan, Song Bai, Lingxi Xie +3
There are two mainstreams for object detection: top-down and bottom-up. The state-of-the-art approaches mostly belong to the first category. In this paper, we demonstrate that the…
Beyond Masking: Demystifying Token-Based Pre-Training for Vision Transformers
Yunjie Tian, Lingxi Xie, Jiemin Fang +6
The past year has witnessed a rapid development of masked image modeling (MIM). MIM is mostly built upon the vision transformers, which suggests that self-supervised visual represe…
MVP: Multimodality-guided Visual Pre-training
Longhui Wei, Lingxi Xie, Wengang Zhou +2
Recently, masked image modeling (MIM) has become a promising direction for visual pre-training. In the context of vision transformers, MIM learns effective visual representation by…
TAPE: Task-Agnostic Prior Embedding for Image Restoration
Lin Liu, Lingxi Xie, Xiaopeng Zhang +5
Learning a generalized prior for natural image restoration is an important yet challenging task. Early methods mostly involved handcrafted priors including normalized sparsity, l_0…
SiamTrans: Zero-Shot Multi-Frame Image Restoration with Pre-Trained Siamese Transformers
Lin Liu, Shanxin Yuan, Jianzhuang Liu +3
We propose a novel zero-shot multi-frame image restoration method for removing unwanted obstruction elements (such as rains, snow, and moire patterns) that vary in successive frame…
Exploring Complicated Search Spaces with Interleaving-Free Sampling
Yunjie Tian, Lingxi Xie, Jiemin Fang +3
The existing neural architecture search algorithms are mostly working on search spaces with short-distance connections. We argue that such designs, though safe and stable, obstacle…
General Greedy De-bias Learning
Xinzhe Han, Shuhui Wang, Chi Su +2
Neural networks often make predictions relying on the spurious correlations from the datasets rather than the intrinsic properties of the task of interest, facing sharp degradation…
NeuSample: Neural Sample Field for Efficient View Synthesis
Jiemin Fang, Lingxi Xie, Xinggang Wang +3
Neural radiance fields (NeRF) have shown great potentials in representing 3D scenes and synthesizing novel views, but the computational overhead of NeRF at the inference stage is s…
Semantic-Aware Generation for Self-Supervised Visual Representation Learning
Yunjie Tian, Lingxi Xie, Xiaopeng Zhang +6
In this paper, we propose a self-supervised visual representation learning approach which involves both generative and discriminative proxies, where we focus on the former part by…
Consensus Synergizes with Memory: A Simple Approach for Anomaly Segmentation in Urban Scenes
Jiazhong Cen, Zenkun Jiang, Lingxi Xie +3
Anomaly segmentation is a crucial task for safety-critical applications, such as autonomous driving in urban scenes, where the goal is to detect out-of-distribution (OOD) objects w…