ICCV2023论文速递(2023.8.14)!共26篇!
最新成果demo展示:
标题:Text2Tex:基于扩散模型的文本驱动纹理合成
论文:https://daveredrum.github.io/Text2Tex/static/Text2Tex.pdf
摘要:
Text2Tex根据给定的文本提示为3D网格生成高质量纹理。我们的方法将修复结合到预先训练的深度感知图像扩散模型中,以从多个视点逐步合成高分辨率部分纹理。为了避免伪影,我们提出了一种自动视图序列生成方案来确定用于更新局部纹理的下一个最佳视图。大量实验表明,我们的方法显著优于现有的文本驱动方法和基于GAN的方法。
ICCV 2023
Updated on : 14 Aug 2023
FunnyBirds: A Synthetic Vision Dataset for a Part-Based Analysis of Explainable AI Methods
论文/Paper: http://arxiv.org/pdf/2308.06248
代码/Code: https://github.com/visinf/funnybirds
Exploring Predicate Visual Context in Detecting of Human-Object Interactions
论文/Paper: http://arxiv.org/pdf/2308.06202
代码/Code: https://github.com/fredzzhang/pvic
RIGID: Recurrent GAN Inversion and Editing of Real Face Videos
论文/Paper: http://arxiv.org/pdf/2308.06097
代码/Code: None
Out-of-Distribution Detection for Monocular Depth Estimation
论文/Paper: http://arxiv.org/pdf/2308.06072
代码/Code: None
Cyclic-Bootstrap Labeling for Weakly Supervised Object Detection
论文/Paper: http://arxiv.org/pdf/2308.05991
代码/Code: https://github.com/yinyf0804/wsod-cbl
Generalizing Event-Based Motion Deblurring in Real-World Scenarios
论文/Paper: http://arxiv.org/pdf/2308.05932
代码/Code: https://github.com/xiangz-0/gem
Collaborative Tracking Learning for Frame-Rate-Insensitive Multi-Object Tracking
论文/Paper: http://arxiv.org/pdf/2308.05911
代码/Code: https://github.com/yolomax/ColTrack
SegDA: Maximum Separable Segment Mask with Pseudo Labels for Domain Adaptive Semantic Segmentation
论文/Paper: http://arxiv.org/pdf/2308.05851
代码/Code: None