CV计算机视觉每日开源代码Paper with code速览-2023.12.6

点击@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：Transformer】Rejuvenating image-GPT as Strong Visual Representation Learners

• 论文地址：https://arxiv.org//pdf/2312.02147

• 开源代码：https://github.com/OliverRensu/D-iGPT

2.【基础网络架构：Transformer】Bootstrapping SparseFormers from Vision Foundation Models

• 论文地址：https://arxiv.org//pdf/2312.01987

• 开源代码：https://github.com/showlab/sparseformer

3.【异常检测】Unsupervised Anomaly Detection using Aggregated Normative Diffusion

• 论文地址：https://arxiv.org//pdf/2312.01904

• 开源代码：https://github.com/alexanderfrotscher/ANDi

4.【视频异常检测】Dynamic Erasing Network Based on Multi-Scale Temporal Features for Weakly Supervised Video Anomaly Detection

• 论文地址：https://arxiv.org//pdf/2312.01764

• 开源代码（即将开源）：https://github.com/ArielZc/DE-Net

5.【图像分割】UniGS: Unified Representation for Image Generation and Segmentation

• 论文地址：https://arxiv.org//pdf/2312.01985

• 开源代码（即将开源）：https://github.com/qqlu/Entity

6.【语义分割】Generalization by Adaptation: Diffusion-Based Domain Extension for Domain-Generalized Semantic Segmentation

• 论文地址：https://arxiv.org//pdf/2312.01850

• 开源代码（即将开源）：https://github.com/JNiemeijer/DIDEX

7.【人脸识别】Effective Adapter for Face Recognition in the Wild

• 论文地址：https://arxiv.org//pdf/2312.01734

• 工程主页：Effective Adapter for Face Recognition in the Wild

• 开源代码（即将开源）：https://github.com/liuyunhaozz/faceadapter/

8.【医学图像分割】MobileUtr: Revisiting the relationship between light-weight CNN and Transformer for efficient medical image segmentation

• 论文地址：https://arxiv.org//pdf/2312.01740

• 开源代码（即将开源）：https://github.com/FengheTan9/MobileUtr

9.【视频超分辨率重建】Motion-Guided Latent Diffusion for Temporally Consistent Real-world Video Super-resolution

• 论文地址：https://arxiv.org//pdf/2312.00853

• 开源代码（即将开源）：https://github.com/IanYeung/MGLD-VSR

10.【图像增强】Enhancing and Adapting in the Clinic: Source-free Unsupervised Domain Adaptation for Medical Image Enhancement

• 论文地址：https://arxiv.org//pdf/2312.01338

• 开源代码：https://github.com/liamheng/Annotation-free-Medical-Image-Enhancement

11.【动作识别】DST-Adapter: Disentangled-and-Deformable Spatio-Temporal Adapter for Few-shot Action Recognition

• 论文地址：https://arxiv.org//pdf/2312.01431

• 开源代码（即将开源）：https://github.com/qizhongtan/D2ST-Adapter

12.【多模态】Aligning and Prompting Everything All at Once for Universal Visual Perception

• 论文地址：https://arxiv.org//pdf/2312.02153

• 开源代码：https://github.com/shenyunhang/APE

13.【多模态】Object Recognition as Next Token Prediction

• 论文地址：https://arxiv.org//pdf/2312.02142

• 开源代码：https://github.com/kaiyuyue/nxtp

14.【多模态】Mitigating Fine-Grained Hallucination by Fine-Tuning Large Vision-Language Models with Caption Rewrites

• 论文地址：https://arxiv.org//pdf/2312.01701

• 开源代码：https://github.com/Anonymousanoy/FOHE

15.【多模态】Good Questions Help Zero-Shot Image Reasoning

• 论文地址：https://arxiv.org//pdf/2312.01598

• 开源代码：https://github.com/kai-wen-yang/QVix

16.【多模态】Unveiling the Power of Audio-Visual Early Fusion Transformers with Dense Interactions through Masked Modeling

• 论文地址：https://arxiv.org//pdf/2312.01017

• 开源代码（即将开源）：https://github.com/stoneMo/DeepAVFusion

17.【多模态】Segment and Caption Anything

• 论文地址：https://arxiv.org//pdf/2312.00869

• 工程主页：Segment and Caption Anything

• 开源代码：https://github.com/xk-huang/segment-caption-anything

18.【多模态】VMC: Video Motion Customization using Temporal Attention Adaption for Text-to-Video Diffusion Models

• 论文地址：https://arxiv.org//pdf/2312.00845

• 工程主页：VMC

• 开源代码：https://github.com/HyeonHo99/Video-Motion-Customization

19.【多模态】A Challenging Multimodal Video Summary: Simultaneously Extracting and Generating Keyframe-Caption Pairs from Video

• 论文地址：https://arxiv.org//pdf/2312.01575

• 开源代码：https://github.com/keitokudo/Multi-VidSum

20.【数字人】GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians

• 论文地址：https://arxiv.org//pdf/2312.02134

• 工程主页：Projectpage of GaussianAvatar

• 开源代码（即将开源）：https://github.com/huliangxiao/GaussianAvatar

21.【数字人】VividTalk: One-Shot Audio-Driven Talking Head Generation Based on 3D Hybrid Prior

• 论文地址：https://arxiv.org//pdf/2312.01841

• 工程主页：VividTalk: One-Shot Audio-Driven Talking Head Generation Based 3D Hybrid Prior

• 开源代码（即将开源）：https://github.com/HumanAIGC/VividTalk

22.【数字人】3DiFACE: Diffusion-based Speech-driven 3D Facial Animation and Editing

• 论文地址：https://arxiv.org//pdf/2312.00870

• 工程主页：3DiFACE: Diffusion-based Speech-driven 3D Facial Animation and Editing

• 开源代码（即将开源）：https://github.com/bala1144/3DiFACE

23.【半监督学习】Virtual Category Learning: A Semi-Supervised Learning Method for Dense Prediction with Extremely Limited Labels

• 论文地址：https://arxiv.org//pdf/2312.01169

• 开源代码：https://github.com/GeoffreyChen777/VC

24.【深度估计】Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation

• 论文地址：https://arxiv.org//pdf/2312.02145

• 工程主页：Marigold: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation

• 开源代码：https://github.com/prs-eth/marigold

25.【深度估计】Deeper into Self-Supervised Monocular Indoor Depth Estimation

• 论文地址：https://arxiv.org//pdf/2312.01283

• 开源代码：https://github.com/fcntes/IndoorDepth

26.【场景补全】PaSCo: Urban 3D Panoptic Scene Completion with Uncertainty Awareness

• 论文地址：https://arxiv.org//pdf/2312.02158

• 工程主页：PaSCo: Urban 3D Panoptic Scene Completion with Uncertainty Awareness

• 开源代码（即将开源）：https://github.com/astra-vision/PaSCo

27.【风格迁移】Multimodality-guided Image Style Transfer using Cross-modal GAN Inversion

• 论文地址：https://arxiv.org//pdf/2312.01671

• 工程主页：Multimodality-guided Image Style Transfer using Cross-modal GAN Inversion

• 代码即将开源

28.【Diffusion】Readout Guidance: Learning Control from Diffusion Features

• 论文地址：https://arxiv.org//pdf/2312.02150

• 工程主页：Readout Guidance: Learning Control from Diffusion Features

• 代码即将开源

29.【Diffusion】ResEnsemble-DDPM: Residual Denoising Diffusion Probabilistic Models for Ensemble Learning

• 论文地址：https://arxiv.org//pdf/2312.01682

• 开源代码（即将开源）：https://github.com/nkicsl/ResEnsemble-DDPM

30.【Diffusion】DeepCache: Accelerating Diffusion Models for Free

• 论文地址：https://arxiv.org//pdf/2312.00858

• 开源代码：https://github.com/horseee/DeepCache

31.【网络剪枝】Visual Prompting Upgrades Neural Network Sparsification: A Data-Model Perspective

• 论文地址：https://arxiv.org//pdf/2312.01397

• 开源代码：https://github.com/UNITES-Lab/VPNs

32.【网络剪枝】Physics Inspired Criterion for Pruning-Quantization Joint Learning

• 论文地址：https://arxiv.org//pdf/2312.00851

• 开源代码：https://github.com/fanxxxxyi/PIC-PQ

33.【姿态估计】Object 6D pose estimation meets zero-shot learning

• 论文地址：https://arxiv.org//pdf/2312.00947

• 工程主页：PoMZ: Object 6D Pose Estimation Meets Zero-Shot Learning

• 代码即将开源

34.【NeRF】Mesh-Guided Neural Implicit Field Editing

• 论文地址：https://arxiv.org//pdf/2312.02157

• 工程主页：Mesh-Guided Neural Implicit Field Editing

• 开源代码（即将开源）：https://github.com/cassiePython/MNeuEdit/tree/master

35.【NeRF】SANeRF-HQ: Segment Anything for NeRF in High Quality

• 论文地址：https://arxiv.org//pdf/2312.01531

• 工程主页：SANeRF-HQ

• 开源代码（即将开源）：https://github.com/lyclyc52/SANeRF-HQ

36.【NeRF】VideoRF: Rendering Dynamic Radiance Fields as 2D Feature Video Streams

• 论文地址：https://arxiv.org//pdf/2312.01407

• 工程主页：VideoRF: Rendering Dynamic Radiance Fields as 2D Feature Video Streams

• 开源代码（即将开源）：https://github.com/aoliao12138/VideoRF

37.【NeRF】Self-Evolving Neural Radiance Fields

• 论文地址：https://arxiv.org//pdf/2312.01003

• 工程主页：SE-NeRF

• 开源代码（即将开源）：https://github.com/KU-CVLAB/SE-NeRF

38.【图像合成】DiffiT: Diffusion Vision Transformers for Image Generation

• 论文地址：https://arxiv.org//pdf/2312.02139

• 开源代码：https://github.com/NVlabs/DiffiT

39.【图像合成】Style Aligned Image Generation via Shared Attention

• 论文地址：https://arxiv.org//pdf/2312.02133

• 工程主页：StyleAlign

• 开源代码：https://github.com/google/style-aligned/

40.【人脸重建】DPHMs: Diffusion Parametric Head Models for Depth-based Tracking

• 论文地址：https://arxiv.org//pdf/2312.01068

• 工程主页：DPHMs: Diffusion Parametric Head Models for Depth-based Tracking

• 开源代码（即将开源）：https://github.com/tangjiapeng/DPHMs

41.【图像检索】Language-only Efficient Training of Zero-shot Composed Image Retrieval

• 论文地址：https://arxiv.org//pdf/2312.01998

• 开源代码：https://github.com/navervision/lincir

42.【Visual Question Answering】How to Configure Good In-Context Sequence for Visual Question Answering

• 论文地址：https://arxiv.org//pdf/2312.01571

• 开源代码：https://github.com/GaryJiajia/OFv2_ICL_VQA

论文已打包，下载链接

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：

CV计算机视觉每日开源代码Paper with code速览-2023.12.5

CV计算机视觉每日开源代码Paper with code速览-2023.12.4

CV计算机视觉每日开源代码Paper with code速览-2023.12.1