ECCV 2024 论文和开源项目合集(Papers with Code)

ECCV 2024 decisions are now available！

注1：欢迎各位大佬提交issue，分享ECCV 2024论文和开源项目！

注2：关于往年CV顶会论文以及其他优质CV论文和大盘点，详见： https://github.com/amusi/daily-paper-computer-vision

• CVPR 2024
• ECCV 2022
• ECCV 2020

想看ECCV 2024和最新最全的顶会工作，欢迎扫码加入【CVer学术交流群】，这是最大的计算机视觉AI知识星球！每日更新，第一时间分享最新最前沿的计算机视觉、深度学习、自动驾驶、医疗影像和AIGC等方向的学习资料，学起来！

【ECCV 2024 论文开源目录】

• 3DGS(Gaussian Splatting)
• Mamba / SSM)
• Avatars
• Backbone
• CLIP
• MAE
• Embodied AI
• GAN
• GNN
• 多模态大语言模型(MLLM)
• 大语言模型(LLM)
• NAS
• OCR
• NeRF
• DETR
• Prompt
• 扩散模型(Diffusion Models)
• ReID(重识别)
• 长尾分布(Long-Tail)
• Vision Transformer
• 视觉和语言(Vision-Language)
• 自监督学习(Self-supervised Learning)
• 数据增强(Data Augmentation)
• 目标检测(Object Detection)
• 异常检测(Anomaly Detection)
• 目标跟踪(Visual Tracking)
• 语义分割(Semantic Segmentation)
• 实例分割(Instance Segmentation)
• 全景分割(Panoptic Segmentation)
• 医学图像(Medical Image)
• 医学图像分割(Medical Image Segmentation)
• 视频目标分割(Video Object Segmentation)
• 视频实例分割(Video Instance Segmentation)
• 参考图像分割(Referring Image Segmentation)
• 图像抠图(Image Matting)
• 图像编辑(Image Editing)
• Low-level Vision
• 超分辨率(Super-Resolution)
• 去噪(Denoising)
• 去模糊(Deblur)
• 自动驾驶(Autonomous Driving)
• 3D点云(3D Point Cloud)
• 3D目标检测(3D Object Detection)
• 3D语义分割(3D Semantic Segmentation)
• 3D目标跟踪(3D Object Tracking)
• 3D语义场景补全(3D Semantic Scene Completion)
• 3D配准(3D Registration)
• 3D人体姿态估计(3D Human Pose Estimation)
• 3D人体Mesh估计(3D Human Mesh Estimation)
• 医学图像(Medical Image)
• 图像生成(Image Generation)
• 视频生成(Video Generation)
• 3D生成(3D Generation)
• 视频理解(Video Understanding)
• 行为识别(Action Recognition)
• 行为检测(Action Detection)
• 文本检测(Text Detection)
• 知识蒸馏(Knowledge Distillation)
• 模型剪枝(Model Pruning)
• 图像压缩(Image Compression)
• 三维重建(3D Reconstruction)
• 深度估计(Depth Estimation)
• 轨迹预测(Trajectory Prediction)
• 车道线检测(Lane Detection)
• 图像描述(Image Captioning)
• 视觉问答(Visual Question Answering)
• 手语识别(Sign Language Recognition)
• 视频预测(Video Prediction)
• 新视点合成(Novel View Synthesis)
• Zero-Shot Learning(零样本学习)
• 立体匹配(Stereo Matching)
• 特征匹配(Feature Matching)
• 场景图生成(Scene Graph Generation)
• 计数(Counting)
• 隐式神经表示(Implicit Neural Representations)
• 图像质量评价(Image Quality Assessment)
• 视频质量评价(Video Quality Assessment)
• 数据集(Datasets)
• 新任务(New Tasks)
• 其他(Others)

<a name="3DGS"></a>

3DGS(Gaussian Splatting)

MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images

• Project: https://donydchen.github.io/mvsplat
• Paper: https://arxiv.org/abs/2403.14627
• Code：https://github.com/donydchen/mvsplat

CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussians

• Paper: https://arxiv.org/abs/2404.01133
• Code: https://github.com/DekuLiuTesla/CityGaussian

FSGS: Real-Time Few-shot View Synthesis using Gaussian Splatting

• Project: https://zehaozhu.github.io/FSGS/
• Paper: https://arxiv.org/abs/2312.00451
• Code: https://github.com/VITA-Group/FSGS

<a name="Mamba"></a>

Mamba / SSM

VideoMamba: State Space Model for Efficient Video Understanding

• Paper: https://arxiv.org/abs/2403.06977
• Code: https://github.com/OpenGVLab/VideoMamba

ZIGMA: A DiT-style Zigzag Mamba Diffusion Model

• Paper: https://arxiv.org/abs/2403.13802
• Code: https://taohu.me/zigma/

<a name="Avatars"></a>

Avatars

<a name="Backbone"></a>

Backbone

<a name="CLIP"></a>

CLIP

<a name="MAE"></a>

MAE

<a name="Embodied-AI"></a>

Embodied AI

<a name="GAN"></a>

GAN

<a name="OCR"></a>

OCR

Bridging Synthetic and Real Worlds for Pre-training Scene Text Detectors

• Paper: https://arxiv.org/pdf/2312.05286

• Code: https://github.com/SJTU-DeepVisionLab/FreeReal

PosFormer: Recognizing Complex Handwritten Mathematical Expression with Position Forest Transformer

• Paper: https://arxiv.org/abs/2407.07764
• Code: https://github.com/SJTU-DeepVisionLab/PosFormer

<a name="Occupancy"></a>

Occupancy

Fully Sparse 3D Occupancy Prediction

• Paper: https://arxiv.org/abs/2312.17118
• Code: https://github.com/MCG-NJU/SparseOcc

<a name="NeRF"></a>

NeRF

NeRF-MAE: Masked AutoEncoders for Self-Supervised 3D Representation Learning for Neural Radiance Fields

• Project: https://nerf-mae.github.io/
• Paper: https://arxiv.org/pdf/2404.01300
• Code: https://github.com/zubair-irshad/NeRF-MAE

<a name="DETR"></a>

DETR

<a name="Prompt"></a>

Prompt

<a name="MLLM"></a>

多模态大语言模型(MLLM)

SQ-LLaVA: Self-Questioning for Large Vision-Language Assistant

• Paper: https://arxiv.org/abs/2403.11299
• Code: https://github.com/heliossun/SQ-LLaVA

ControlCap: Controllable Region-level Captioning

• Paper: https://arxiv.org/abs/2401.17910
• Code: https://github.com/callsys/ControlCap

<a name="LLM"></a>

大语言模型(LLM)

<a name="NAS"></a>

NAS

<a name="ReID"></a>

ReID(重识别)

<a name="Diffusion"></a>

扩散模型(Diffusion Models)

ZIGMA: A DiT-style Zigzag Mamba Diffusion Model

• Paper: https://arxiv.org/abs/2403.13802
• Code: https://taohu.me/zigma/

Skews in the Phenomenon Space Hinder Generalization in Text-to-Image Generation

• Paper: https://arxiv.org/abs/2403.16394
• Code: https://github.com/zdxdsw/skewed_relations_T2I

The Lottery Ticket Hypothesis in Denoising: Towards Semantic-Driven Initialization

• Project: https://ut-mao.github.io/noise.github.io/
• Paper: https://arxiv.org/abs/2312.08872
• Code: https://github.com/UT-Mao/Initial-Noise-Construction

<a name="Vision-Transformer"></a>

Vision Transformer

GiT: Towards Generalist Vision Transformer through Universal Language Interface

• Paper: https://arxiv.org/abs/2403.09394
• Code: https://github.com/Haiyang-W/GiT

<a name="VL"></a>

视觉和语言(Vision-Language)

GalLoP: Learning Global and Local Prompts for Vision-Language Models

• Paper：https://arxiv.org/abs/2407.01400

<a name="Object-Detection"></a>

目标检测(Object Detection)

Relation DETR: Exploring Explicit Position Relation Prior for Object Detection

• Paper: https://arxiv.org/abs/2407.11699v1
• Code: https://github.com/xiuqhou/Relation-DETR
• Dataset: https://huggingface.co/datasets/xiuqhou/SA-Det-100k

Cross-Domain Few-Shot Object Detection via Enhanced Open-Set Object Detector

• Project: http://yuqianfu.com/CDFSOD-benchmark/
• Paper: https://arxiv.org/pdf/2402.03094
• Code: https://github.com/lovelyqian/CDFSOD-benchmark

<a name="Anomaly-Detection"></a>

异常检测(Anomaly Detection)

<a name="VT"></a>

目标跟踪(Object Tracking)

<a name="Semantic-Segmentation"></a>

语义分割(Semantic Segmentation)

Context-Guided Spatial Feature Reconstruction for Efficient Semantic Segmentation

• Paper: https://arxiv.org/abs/2405.06228

• Code: https://github.com/nizhenliang/CGRSeg

<a name="MI"></a>

医学图像(Medical Image)

Brain-ID: Learning Contrast-agnostic Anatomical Representations for Brain Imaging

• Paper: https://arxiv.org/abs/2311.16914
• Code: https://github.com/peirong26/Brain-ID

FairDomain: Achieving Fairness in Cross-Domain Medical Image Segmentation and Classification

• Project: https://ophai.hms.harvard.edu/datasets/harvard-fairdomain20k
• Paper : https://arxiv.org/abs/2407.08813
• Dataset: https://drive.google.com/drive/u/1/folders/1huH93JVeXMj9rK6p1OZRub868vv0UK0O
• Code: https://github.com/Harvard-Ophthalmology-AI-Lab/FairDomain

<a name="MIS"></a>

医学图像分割(Medical Image Segmentation)

ScribblePrompt: Fast and Flexible Interactive Segmentation for Any Biomedical Image

• Project: https://scribbleprompt.csail.mit.edu/
• Paper: https://arxiv.org/abs/2312.07381
• Code: https://github.com/halleewong/ScribblePrompt

AnatoMask: Enhancing Medical Image Segmentation with Reconstruction-guided Self-masking

• Paper: https://arxiv.org/abs/2407.06468
• Code: https://github.com/ricklisz/AnatoMask

Representing Topological Self-Similarity Using Fractal Feature Maps for Accurate Segmentation of Tubular Structures

• Paper: https://arxiv.org/abs/2407.14754
• Code: https://github.com/cbmi-group/FFM-Multi-Decoder-Network

<a name="VOS"></a>

视频目标分割(Video Object Segmentation)

DVIS-DAQ: Improving Video Segmentation via Dynamic Anchor Queries

• Project: https://zhang-tao-whu.github.io/projects/DVIS_DAQ/
• Paper: https://arxiv.org/abs/2404.00086
• Code: https://github.com/zhang-tao-whu/DVIS_Plus

<a name="Autonomous-Driving"></a>

自动驾驶(Autonomous Driving)

Fully Sparse 3D Occupancy Prediction

• Paper: https://arxiv.org/abs/2312.17118
• Code: https://github.com/MCG-NJU/SparseOcc

milliFlow: Scene Flow Estimation on mmWave Radar Point Cloud for Human Motion Sensing

• 论文: https://arxiv.org/abs/2306.17010
• 代码: https://github.com/Toytiny/milliFlow/

4D对比超流是密集3D表征学习器

• 论文: https://arxiv.org/abs/2407.06190
• 代码: https://github.com/Xiangxu-0103/SuperFlow

<a name="3D-Point-Cloud"></a>

3D点云(3D-Point-Cloud)

<a name="3DOD"></a>

3D目标检测(3D Object Detection)

具有动态空间剪枝的3D小目标检测

• 项目: https://xuxw98.github.io/DSPDet3D/
• 论文: https://arxiv.org/abs/2305.03716
• 代码: https://github.com/xuxw98/DSPDet3D

光线去噪：基于深度感知的难负样本采样用于多视图3D目标检测

• 论文: https://arxiv.org/abs/2402.03634
• 代码: https://github.com/LiewFeng/RayDN

<a name="3DOD"></a>

3D语义分割(3D Semantic Segmentation)

<a name="Image-Editing"></a>

图像编辑(Image Editing)

<a name="Image-Inpainting"></a>

图像补全/图像修复(Image Inpainting)

BrushNet：具有分解双分支扩散的即插即用图像补全模型

• 项目: https://tencentarc.github.io/BrushNet/
• 论文: https://arxiv.org/abs/2403.06976
• 代码: https://github.com/TencentARC/BrushNet

<a name="Video-Editing"></a>

视频编辑(Video Editing)

<a name="LLV"></a>

低级视觉(Low-level Vision)

通过直方图变换在恶劣天气条件下恢复图像

• 论文: https://arxiv.org/abs/2407.10172
• 代码: https://github.com/sunshangquan/Histoformer

OneRestore：复合退化的通用恢复框架

• 项目: https://gy65896.github.io/projects/ECCV2024_OneRestore
• 论文: https://arxiv.org/abs/2407.04621
• 代码: https://github.com/gy65896/OneRestore

超分辨率(Super-Resolution)

<a name="Denoising"></a>

去噪(Denoising)

图像去噪(Image Denoising)

<a name="3D-Human-Pose-Estimation"></a>

3D人体姿态估计(3D Human Pose Estimation)

<a name="Image-Generation"></a>

图像生成(Image Generation)

在文本到图像扩散模型中基于对象条件的能量式注意力图对齐

• 论文: https://arxiv.org/abs/2404.07389
• 代码: https://github.com/YasminZhang/EBAMA

每个像素都有其时刻：通过密集归一化实现超高分辨率的无对齐图像到图像翻译

• 项目: https://kaminyou.com/Dense-Normalization/
• 论文: https://arxiv.org/abs/2407.04245
• 代码: https://github.com/Kaminyou/Dense-Normalization

ZIGMA：一种DiT风格的之字形曼巴扩散模型

• 论文: https://arxiv.org/abs/2403.13802
• 代码: https://taohu.me/zigma/

现象空间中的倾斜阻碍了文本到图像生成的泛化

• 论文: https://arxiv.org/abs/2403.16394
• 代码: https://github.com/zdxdsw/skewed_relations_T2I

<a name="Video-Generation"></a>

视频生成(Video Generation)

VideoStudio：生成内容一致和多场景的视频

• 项目: https://vidstudio.github.io/
• 代码: https://github.com/FuchenUSTC/VideoStudio

<a name="3D-Generation"></a>

3D生成

<a name="Video-Understanding"></a>

视频理解(Video Understanding)

VideoMamba：高效视频理解的状态空间模型

• 论文: https://arxiv.org/abs/2403.06977
• 代码: https://github.com/OpenGVLab/VideoMamba

C2C：用于零样本组合动作识别的组件到组合学习

• 论文: https://arxiv.org/abs/2407.06113
• 代码: https://github.com/RongchangLi/ZSCAR_C2C

<a name="Action-Recognition"></a>

行为识别(Action Recognition)

SA-DVAE：通过解耦变分自编码器改进零样本基于骨架的动作识别

• 论文: https://arxiv.org/abs/2407.13460
• 代码: https://github.com/pha123661/SA-DVAE

<a name="KD"></a>

知识蒸馏(Knowledge Distillation)

<a name="IC"></a>

图像压缩(Image Compression)

基于空间频率自适应的机器和人类视觉图像压缩

• 代码: https://github.com/qingshi9974/ECCV2024-AdpatICMH
• 论文: http://arxiv.org/abs/2407.09853

<a name="Stereo-Matching"></a>

立体匹配(Stereo Matching)

<a name="SGG"></a>

场景图生成(Scene Graph Generation)

<a name="Counting"></a>

计数(Counting)

通过良好示例进行零样本目标计数

• 论文: https://arxiv.org/abs/2407.04948
• 代码: https://github.com/HopooLinZ/VA-Count

<a name="Video-Quality-Assessment"></a>

视频质量评价(Video Quality Assessment)

<a name="Datasets"></a>

数据集(Datasets)

其他(Others)

用于3D视觉定位的多分支协作学习网络

• 论文: https://arxiv.org/abs/2407.05363v2
• 代码: https://github.com/qzp2018/MCLN

PDiscoFormer：通过视觉变压器放宽部分发现约束

• 代码: https://github.com/ananthu-aniraj/pdiscoformer
• 论文: https://arxiv.org/abs/2407.04538

SPVLoc：在未知环境中用于6D相机定位的语义全景视口匹配

• 项目: https://fraunhoferhhi.github.io/spvloc/
• 论文: https://arxiv.org/abs/2404.10527
• 代码: https://github.com/fraunhoferhhi/spvloc

REFRAME：移动设备实时渲染反射表面

• 项目: https://xdimlab.github.io/REFRAME/
• 论文: https://arxiv.org/abs/2403.16481
• 代码: https://github.com/MARVELOUSJI/REFRAME