CV计算机视觉每日开源代码Paper with code速览-2023.10.13

精华置顶

墙裂推荐！小白如何1个月系统学习CV核心知识：链接

点击@CV计算机视觉，关注更多CV干货

论文已打包，点击进入—>下载界面

点击加入—>CV计算机视觉交流群

1.【基础网络架构】CHIP: Contrastive Hierarchical Image Pretraining

论文地址：https://arxiv.org//pdf/2310.08304
开源代码：GitHub - harshiljhaveri/CHIP

2.【基础网络架构：Transformer】AutoVP: An Automated Visual Prompting Framework and Benchmark

论文地址：https://arxiv.org//pdf/2310.08381
开源代码：GitHub - IBM/AutoVP: Code and Benchmark for the paper "AutoVP: An Automated Visual Prompting Framework and Benchmark"

3.【关键点检测】UniPose: Detecting Any Keypoints

论文地址：https://arxiv.org//pdf/2310.08530
工程主页：UniPose: Detecting Any Keypoints
开源代码（即将开源）：GitHub - IDEA-Research/UniPose: Official implementation of the paper "UniPose : Detecting Any Keypoints"

4.【点云】PonderV2: Pave the Way for 3D Foundataion Model with A Universal Pre-training Paradigm

论文地址：https://arxiv.org//pdf/2310.08586
开源代码：GitHub - Pointcept/Pointcept: Pointcept: a codebase for point cloud perception research. Latest works: PPT, MSC (CVPR'23), PTv2 (NeurIPS'22)

5.【点云分割】PointHR: Exploring High-Resolution Architectures for 3D Point Cloud Segmentation

论文地址：https://arxiv.org//pdf/2310.07743
开源代码：GitHub - haibo-qiu/PointHR: PointHR: Exploring High-Resolution Architectures for 3D Point Cloud Segmentation

6.【医学图像分割】Volumetric Medical Image Segmentation via Scribble Annotations and Shape Priors

论文地址：https://arxiv.org//pdf/2310.08084
开源代码：GitHub - Qybc/Scribble2D5: Scribble2D5: Weakly-Supervised Volumetric Image Segmentation via Scribble Annotations

7.【医学图像分割：3D】3D TransUNet: Advancing Medical Image Segmentation through Vision Transformers

论文地址：https://arxiv.org//pdf/2310.07781
开源代码：GitHub - Beckschen/3D-TransUNet: This is the official repository for the paper "3D TransUNet: Advancing Medical Image Segmentation through Vision Transformers"

8.【多模态】Visual Data-Type Understanding does not emerge from Scaling Vision-Language Models

论文地址：https://arxiv.org//pdf/2310.08577
开源代码（即将开源）：GitHub - bethgelab/DataTypeIdentification: Code for the paper: "Visual Data-Type Understanding does not emerge from Scaling Vision-Language Models"

9.【多模态】Multimodal Variational Auto-encoder based Audio-Visual Segmentation

论文地址：https://arxiv.org//pdf/2310.08303
工程主页：Multimodal Variational Auto-encoder based Audio-Visual Segmentation
开源代码（即将开源）：GitHub - OpenNLPLab/MMVAE-AVS: Multimodal Variational Auto-encoder based Audio-Visual Segmentation [ICCV2023].

10.【多模态】Distilling from Vision-Language Models for Improved OOD Generalization in Vision Tasks

论文地址：https://arxiv.org//pdf/2310.08255
开源代码：GitHub - val-iisc/VL2V-ADiP: Distilling from Vision-Language Models for Improved OOD Generalization in Image Classification

11.【多模态】Lifelong Audio-video Masked Autoencoder with Forget-robust Localized Alignments

论文地址：https://arxiv.org//pdf/2310.08204
工程主页：FLAVA
代码即将开源

12.【多模态】Generalized Logit Adjustment: Calibrating Fine-tuned Models by Removing Label Bias in Foundation Models

论文地址：https://arxiv.org//pdf/2310.08106
开源代码（即将开源）：GitHub - BeierZhu/GLA: [NeurIPS 2023] Generalized Logit Adjustment (Coming Soon)

13.【多模态】SingleInsert: Inserting New Concepts from a Single Image into Text-to-Image Models for Flexible Editing

论文地址：https://arxiv.org//pdf/2310.08094
工程主页：SingleInsert
开源代码（即将开源）：GitHub - JarrentWu1031/SingleInsert: Official pytorch implementation for SingleInsert

14.【多模态】Can We Edit Multimodal Large Language Models?

论文地址：https://arxiv.org//pdf/2310.08475
开源代码：GitHub - zjunlp/EasyEdit: An Easy-to-use Knowledge Editing Framework for LLMs.

15.【自动驾驶：多模态感知】UniPAD: A Universal Pre-training Paradigm for Autonomous Driving

论文地址：https://arxiv.org//pdf/2310.08370
开源代码（即将开源）：GitHub - Nightmare-n/UniPAD: UniPAD: A Universal Pre-training Paradigm for Autonomous Driving

16.【自动驾驶：协同感知】DUSA: Decoupled Unsupervised Sim2Real Adaptation for Vehicle-to-Everything Collaborative Perception

论文地址：https://arxiv.org//pdf/2310.08117
开源代码（即将开源）：GitHub - refkxh/DUSA: [ACM MM 2023] Official implementation of DUSA: Decoupled Unsupervised Sim2Real Adaptation for Vehicle-to-Everything Collaborative Perception

17.【自动驾驶：仿真】DrivingDiffusion: Layout-Guided multi-view driving scene video generation with latent diffusion model

论文地址：https://arxiv.org//pdf/2310.07771
工程主页：DrivingDiffusion: Layout-Guided multi-view driving scene video generation with latent diffusion model
开源代码（即将开源）：GitHub - shalfun/DrivingDiffusion: Layout-Guided multi-view driving scene video generation with latent diffusion model

18.【Diffusion】HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion

论文地址：https://arxiv.org//pdf/2310.08579
工程主页：HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion
开源代码（即将开源）：GitHub - snap-research/HyperHuman: Github Repo for "HyperHuman: Hyper-Realistic Human Generation with Latent Structural Diffusion"

19.【Diffusion】MotionDirector: Motion Customization of Text-to-Video Diffusion Models

论文地址：https://arxiv.org//pdf/2310.08465
代码即将开源

20.【人体姿态估计】X-HRNet: Towards Lightweight Human Pose Estimation with Spatially Unidimensional Self-Attention

论文地址：https://arxiv.org//pdf/2310.08042
开源代码：GitHub - cool-xuan/x-hrnet: Official code for "X-HRNet: Towards Lightweight Human Pose Estimation with Spatially Unidimensional Self-Attention"

21.【人体运动生成】OmniControl: Control Any Joint at Any Time for Human Motion Generation

论文地址：https://arxiv.org//pdf/2310.08580
工程主页：OmniControl
开源代码（即将开源）：GitHub - neu-vi/OmniControl

22.【生成模型】Explorable Mesh Deformation Subspaces from Unstructured Generative Models

论文地址：https://arxiv.org//pdf/2310.07814
开源代码（即将开源）：ArmanMaesumi/generative-mesh-subspaces · GitHub

23.【三维重建】Consistent123: Improve Consistency for One Image to 3D Object Synthesis

论文地址：https://arxiv.org//pdf/2310.08092
工程主页：Consistent123: Improve Consistency for One Image to 3D Object Synthesis
代码即将开源

24.【图像分类：长尾分布】Long-Tailed Classification Based on Coarse-Grained Leading Forest and Multi-Center Loss

论文地址：https://arxiv.org//pdf/2310.08206
开源代码（即将开源）：GitHub - jinyery/Cognisance: Long-tail Classification Based on Invariant Feature Learning from A Multi-granularity Perspective

论文已打包，点击进入—>下载界面

CV计算机视觉交流群

群内包含目标检测、图像分割、目标跟踪、Transformer、多模态、NeRF、GAN、缺陷检测、显著目标检测、关键点检测、超分辨率重建、SLAM、人脸、OCR、生物医学图像、三维重建、姿态估计、自动驾驶感知、深度估计、视频理解、行为识别、图像去雾、图像去雨、图像修复、图像检索、车道线检测、点云目标检测、点云分割、图像压缩、运动预测、神经网络量化、网络部署等多个领域的大佬，不定期分享技术知识、面试技巧和内推招聘信息。

想进群的同学请添加微信号联系管理员：PingShanHai666。添加好友时请备注：学校/公司+研究方向+昵称。

推荐阅读：

HSN：微调预训练ViT用于目标检测和语义分割，华南理工和阿里巴巴联合提出

EViT：借鉴鹰眼视觉结构，南开大学等提出ViT新骨干架构，在多个任务上涨点

CV计算机视觉每日开源代码Paper with code速览-2023.10.12

CV计算机视觉每日开源代码Paper with code速览-2023.10.11

CV计算机视觉每日开源代码Paper with code速览-2023.10.10