计算机视觉开源代码汇总

1.【基础网络架构】Regularization of polynomial networks for image recognition

论文地址：https://arxiv.org/pdf/2303.13896.pdf

开源代码:https://github.com/grigorisg9gr/regularized_polynomials

2.【目标检测：域自适应】2PCNet: Two-Phase Consistency Training for Day-to-Night Unsupervised Domain Adaptive Object Detection

论文地址: https://arxiv.org/pdf/2303.13853.pdf

开源代码:https://github.com/mecarill/2pcnet

4.【目标跟踪：数据集】ARKitTrack: A New Diverse Dataset for Tracking Using Mobile RGB-D Data

论文地址:**https://(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13885.pdf)**

* 工程主页：https://arkittrack.github.io/

* 开源代码（即将开源）:GitHub - lawrence-cj/ARKitTrack: PyTorch implementation of ARKitTrack for CVPR'2023 paper "ARKitTrack: A New Diverse Dataset for Tracking Using Mobile RGB-D Data", by Haojie Zhao, Junsong Chen, Lijun Wang, Huchuan Lu. Code will be released here.(https://github.com/lawrence-cj/ARKitTrack)

5.【异常检测】Anomaly Detection under Distribution Shift

* 论文地址:**https://arxiv.org/pdf/2303.13845.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13845.pdf)**

* 开源代码（即将开源）:**https://github.com/mala-lab/ADS(https://link.zhihu.com/?target=https%3A//github.com/mala-lab/ADShift)**

7.【视觉3D目标检测】MoGDE: Boosting Mobile Monocular 3D Object Detection with Ground Depth Estimation

* 论文地址：**https://arxiv.org/pdf/2303.13561.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13561.pdf)**

* 代码即将开源

8.【3D目标检测】Collaboration Helps Camera Overtake LiDAR in 3D Detection

* 论文地址：**https://arxiv.org/pdf/2303.13560.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13560.pdf)**

* 开源代码（即将开源）：**https://github.com/MediaBrain-S(https://link.zhihu.com/?target=https%3A//github.com/MediaBrain-SJTU/CoCa3D)**

正在上传...重新上传取消

9.【医学图像分割:半监督】Inherent Consistent Learning for Accurate Semi-supervised Medical Image Segmentation

* 论文地址：**https://arxiv.org/pdf/2303.14175.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.14175.pdf)**

* 开源代码:**https://github.com/zhuye98/ICL.(https://link.zhihu.com/?target=https%3A//github.com/zhuye98/ICL.git)**

正在上传...重新上传取消

10.【医学图像分割：Few Shot】Few Shot Medical Image Segmentation with Cross Attention Transformer

* 论文地址：**https://arxiv.org/pdf/2303.13867.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13867.pdf)**

* 代码即将开源

11.【三维重建】BundleSDF: Neural 6-DoF Tracking and 3D Reconstruction of Unknown Objects

* 论文地址:**https://arxiv.org/pdf/2303.14158.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.14158.pdf)**

* 工程主页：**https://bundlesdf.github.io/(https://link.zhihu.com/?target=https%3A//bundlesdf.github.io/)**

* 代码即将开源

12.【人脸重建】NeuFace: Realistic 3D Neural Face Rendering from Multi-view Images

* 论文地址：**https://arxiv.org/pdf/2303.14092.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.14092.pdf)**

* 开源代码：**https://github.com/aejion/NeuFace(https://link.zhihu.com/?target=https%3A//github.com/aejion/NeuFace)**

13.【类别增量学习】Class-Incremental Exemplar Compression for Class-Incremental Learning

* 论文地址：**https://arxiv.org/pdf/2303.14042.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.14042.pdf)**

* 开源代码（即将开源）：**https://github.com/xfflzl/CIM-C(https://link.zhihu.com/?target=https%3A//github.com/xfflzl/CIM-CIL)**

14.【自动驾驶：场景补全】StereoScene: BEV-Assisted Stereo Matching Empowers 3D Semantic Scene Completion

* 论文地址：**https://arxiv.org/pdf/2303.13959.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13959.pdf)**

* 开源代码：**https://github.com/Arlo0o/Stere(https://link.zhihu.com/?target=https%3A//github.com/Arlo0o/StereoScene)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-7055375cffd8f3ab.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

!image.png(https://upload-images.jianshu.io/upload_images/18639252-305205771b4acacb.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

15.【类别增量学习】Two-level Graph Network for Few-Shot Class-Incremental Learning

* 论文地址：**https://arxiv.org/pdf/2303.13862.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13862.pdf)**

* 开源代码（即将开源）：**https://github.com/sukechenhao/(https://link.zhihu.com/?target=https%3A//github.com/sukechenhao/SCGN)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-32d38fe1cc32b4db.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

16.【神经网络量化】Hard Sample Matters a Lot in Zero-Shot Quantization

* 论文地址：**https://arxiv.org/pdf/2303.13826.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13826.pdf)**

* 开源代码：**https://github.com/lihuantong/HAST(https://link.zhihu.com/?target=https%3A//github.com/lihuantong/HAST)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-f6af27639c82f5df.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

20.【姿态估计】NOPE: Novel Object Pose Estimation from a Single Image

* 论文地址:**https://arxiv.org/pdf/2303.13612.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13612.pdf)**

* 开源代码（即将开源）：**https://github.com/nv-nguyen/nope(https://link.zhihu.com/?target=https%3A//github.com/nv-nguyen/nope)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-1ecc94a9182edd54.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

21.【医学图像】Leveraging Old Knowledge to Continually Learn New Classes in Medical Images

* 论文地址:**https://arxiv.org/pdf/2303.13752.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13752.pdf)**

* 开源代码:**https://github.com/EvelynChee/LO2LN(https://link.zhihu.com/?target=https%3A//github.com/EvelynChee/LO2LN)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-b2998e6205619f2e.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

22.【点云异常检测】Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection

* 论文地址：**https://arxiv.org/ftp/arxiv/papers/2303/2303.13194.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/ftp/arxiv/papers/2303/2303.13194.pdf)**

* 开源代码（即将开源）：**https://github.com/caoyunkang/CPMF(https://link.zhihu.com/?target=https%3A//github.com/caoyunkang/CPMF)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-f3cdcf9ced296900.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

18.【点云分割】Position-Guided Point Cloud Panoptic Segmentation Transformer

* 论文地址：**https://arxiv.org/pdf/2303.13509.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13509.pdf)**

* 开源代码（即将开源）：**https://github.com/SmartBot-PJLab/P3Former(https://link.zhihu.com/?target=https%3A//github.com/SmartBot-PJLab/P3Former)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-679138cbf3b9874a.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

17.【点云3D目标检测：自监督预训练】MV-JAR: Masked Voxel Jigsaw and Reconstruction for LiDAR-Based Self-Supervised Pre-Training

* 论文地址：**https://arxiv.org/pdf/2303.13510.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13510.pdf)**

* 开源代码（即将开源）：**https://github.com/SmartBot-PJL(https://link.zhihu.com/?target=https%3A//github.com/SmartBot-PJLab/MV-JAR)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-2ca7ba07786c9809.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

15.【点云：自监督学习】PointGame: Geometrically and Adaptively Masked Auto-Encoder on Point Clouds

* 论文地址：**https://arxiv.org/pdf/2303.1310(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13100.pdf)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-9af4031dcef1d9ca.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

7.【异常检测：医学图像】Confidence-Aware and Self-Supervised Image Anomaly Localisation

* 论文地址：**https://arxiv.org/pdf/2303.13227.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13227.pdf)**

* 代码即将开源

!image.png(https://upload-images.jianshu.io/upload_images/18639252-c07081abdd2ea080.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)

2.【动作识别】A Large-scale Study of Spatiotemporal Representation Learning with a New Benchmark on Action Recognition

* 论文地址：**https://arxiv.org/pdf/2303.13505.pdf(https://link.zhihu.com/?target=https%3A//arxiv.org/pdf/2303.13505.pdf)**

* 开源代码：**https://github.com/AndongDeng/B(https://link.zhihu.com/?target=https%3A//github.com/AndongDeng/BEAR)**

!image.png(https://upload-images.jianshu.io/upload_images/18639252-b82c91d967467220.png?imageMogr2/auto-orient/strip|imageView2/2/w/1240)