Update ICCV2021.md

ICCV2021.md

@@ -254,6 +254,20 @@
 [1] Unsupervised Domain Adaptive 3D Detection with Multi-Level Consistency<br>
 [paper](https://arxiv.org/pdf/2107.11355.pdf)<br><br>
 
+<a name="VOD"/>
+
+### 视频目标检测(Video Object Detection)
+
+[1] Social Fabric: Tubelet Compositions for Video Relation Detection<br>
+[paper](https://arxiv.org/abs/2108.08363) | [code](https://github.com/shanshuo/Social-Fabric)<br><br>
+
+<a name="HOI"/> 
+
+### 人物交互检测(HOI Detection)
+
+[1] Exploiting Scene Graphs for Human-Object Interaction Detection<br>
+[paper](https://arxiv.org/abs/2108.08584) | [code](https://github.com/ht014/SG2HOI)<br><br>
+
 <a name="SOD"/> 
 
 ### 显著性目标检测(Saliency Object Detection)
@@ -422,6 +436,9 @@
 
 ### 人脸识别/检测(Facial Recognition/Detection)
 
+[3] Understanding and Mitigating Annotation Bias in Facial Expression Recognition<br>
+[paper](https://arxiv.org/abs/2108.08504)<br><br>
+
 [2] SynFace: Face Recognition with Synthetic Data<br>
 [paper](https://arxiv.org/abs/2108.07960)<br><br>
 
@@ -475,6 +492,9 @@
 
 ### 点云(Point Cloud)
 
+[11] PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers(点云补全)(Oral)<br>
+[paper](https://arxiv.org/abs/2108.08839) | [code](https://github.com/yuxumin/PoinTr)<br><br>
+
 [10] ME-PCN: Point Completion Conditioned on Mask Emptiness(点云补全)<br>
 [paper](https://arxiv.org/abs/2108.08187)<br><br>
 
@@ -509,6 +529,21 @@
 
 ### 三维重建(3D Reconstruction)
 
+[11] Gravity-Aware Monocular 3D Human-Object Reconstruction<br>
+[paper](https://arxiv.org/abs/2108.08844) | [code](http://4dqv.mpi-inf.mpg.de/GraviCap/)<br><br>
+
+[10] 3DIAS: 3D Shape Reconstruction with Implicit Algebraic Surfaces<br>
+[paper](https://arxiv.org/abs/2108.08653) | [code](https://myavartanoo.github.io/3dias/)<br><br>
+
+[9] VolumeFusion: Deep Depth Fusion for 3D Scene Reconstruction<br>
+[paper](https://arxiv.org/abs/2108.08623)<br><br>
+
+[8] Learning Anchored Unsigned Distance Functions with Gradient Direction Alignment for Single-view Garment Reconstruction<br>
+[paper](https://arxiv.org/abs/2108.08478)<br><br>
+
+[7] Vis2Mesh: Efficient Mesh Reconstruction from Unstructured Point Clouds of Large Scenes with Learned Virtual View Visibility<br>
+[paper](https://arxiv.org/abs/2108.08378) | [code](https://github.com/GDAOSU/vis2mesh)<br>
+
 [6] Deep Hybrid Self-Prior for Full 3D Mesh Generation<br>
 [paper](https://arxiv.org/abs/2108.08017) | [project](https://yqdch.github.io/DHSP3D)<br><br>
 
@@ -557,6 +592,12 @@
 
 ### Attention
 
+[6] Causal Attention for Unbiased Visual Recognition<br>
+[paper](https://arxiv.org/abs/2108.08782) | [code](https://github.com/Wangt-CN/CaaM)<br><br>
+
+[5] Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification(因果推理)(细粒度识别)<br>
+[paper](https://arxiv.org/abs/2108.08728) | [code](https://github.com/raoyongming/CAL)<br><br>
+
 [4] Residual Attention: A Simple but Effective Method for Multi-Label Recognition<br>
 [paper](https://arxiv.org/abs/2108.02456)<br><br>
 
@@ -763,6 +804,12 @@
 
 ## GAN/生成式/对抗式(GAN/Generative/Adversarial)
 
+[15] Towards Vivid and Diverse Image Colorization with Generative Color Prior(图像着色)<br>
+[paper](https://arxiv.org/abs/2108.08826)<br><br>
+
+[14] Exploiting Multi-Object Relationships for Detecting Adversarial Attacks in Complex Scenes<br>
+[paper](https://arxiv.org/abs/2108.08421)<br><br>
+
 [13] Unsupervised Geodesic-preserved Generative Adversarial Networks for Unconstrained 3D Pose Transfer<br>
 [paper](https://arxiv.org/abs/2108.07520) ｜ [code](https://github.com/mikecheninoulu/Unsupervised_IEPGAN)<br><br>
 
@@ -811,6 +858,9 @@
 
 ## 图像处理(Image Processing)
 
+[4] Towards Vivid and Diverse Image Colorization with Generative Color Prior(图像着色)<br>
+[paper](https://arxiv.org/abs/2108.08826)<br><br>
+
 [3] Hierarchical Conditional Flow: A Unified Framework for Image Super-Resolution and Image Rescaling<br>
 [paper](https://arxiv.org/abs/2108.05301) | [code](https://github.com/JingyunLiang/HCFlow)<br><br>
 
@@ -830,6 +880,16 @@
 [1] Learning for Scale-Arbitrary Super-Resolution from Scale-Specific Networks<br>
 [paper](https://arxiv.org/abs/2004.03791) | [code](https://github.com/LongguangWang/ArbSR)<br><br>
 
+<a name="ImageRestoration"/> 
+
+### 图像复原/图像增强(Image Restoration)
+
+[2] Real-time Image Enhancer via Learnable Spatial-aware 3D Lookup Tables<br>
+[paper](https://arxiv.org/abs/2108.08697)<br><br>
+
+[1] Spatially-Adaptive Image Restoration using Distortion-Guided Networks<br>
+[paper](https://arxiv.org/abs/2108.08617) | [code](https://github.com/human-analysis/spatially-adaptive-image-restoration)<br><br>
+
 <a name="ImageDenoising"/> 
 
 ### 图像去噪/去模糊/去雨去雾(Image Denoising)
@@ -880,6 +940,9 @@
 
 ### 姿态估计(Human Pose Estimation)
 
+[9] DECA: Deep viewpoint-Equivariant human pose estimation using Capsule Autoencoders(Oral)<br>
+[paper](https://arxiv.org/abs/2108.08557) | [code](https://github.com/mmlab-cv/DECA)<br><br>
+
 [8] Learning Skeletal Graph Neural Networks for Hard 3D Pose Estimation<br>
 [paper](https://arxiv.org/abs/2108.07181) | [code](https://github.com/ailingzengzzz/Skeletal-GNN)<br><br>
 
@@ -904,10 +967,23 @@
 [1] HuMoR: 3D Human Motion Model for Robust Pose Estimation(Oral)<br>
 [paper](https://geometry.stanford.edu/projects/humor/docs/humor.pdf) | [video](https://youtu.be/5VWirxUHG0Y) | [project](https://geometry.stanford.edu/projects/humor/)<br><br>
 
+<a name="Flow/Pose/MotionEstimation"/> 
+
+### 光流/位姿/运动估计(Flow/Pose/Motion Estimation)
+
+[1] SO-Pose: Exploiting Self-Occlusion for Direct 6D Pose Estimation<br>
+[paper](https://arxiv.org/abs/2108.08367)<br><br>
+
 <a name="DepthEstimation"/> 
 
 ### 深度估计(Depth Estimation)
 
+[6] Fine-grained Semantics-aware Representation Enhancement for Self-supervised Monocular Depth Estimation(Oral)<br>
+[paper](https://arxiv.org/abs/2108.08829) | [code](https://github.com/hyBlue/FSRE-Depth)<br><br>
+
+[5] StructDepth: Leveraging the structural regularities for self-supervised indoor depth estimation<br>
+[paper](https://arxiv.org/abs/2108.08574) | [code](https://github.com/SJTU-ViSYS/StructDepth)<br><br>
+
 [4] Self-supervised Monocular Depth Estimation for All Day Images using Domain Separation<br>
 [paper](https://arxiv.org/abs/2108.07628)<br><br>
 
@@ -925,6 +1001,9 @@
 
 ## 图像&视频检索/理解(Image&Video Retrieval/Video Understanding)
 
+[6] Universal Cross-Domain Retrieval: Generalizing Across Classes and Domains<br>
+[paper](https://arxiv.org/abs/2108.08356)<br><br>
+
 [5] ASMR: Learning Attribute-Based Person Search with Adaptive Semantic Margin Regularizer<br>
 [paper](https://arxiv.org/abs/2108.04533)<br><br>
 
@@ -1017,6 +1096,8 @@
 
 ### 图像匹配(Image Matching)
 
+[6] Learning to Match Features with Seeded Graph Matching Network<br>
+[paper](https://arxiv.org/abs/2108.08771) | [code](https://github.com/vdvchen/SGMNet)<br><br>
 
 [5] Pixel-Perfect Structure-from-Motion with Featuremetric Refinement<br>
 [paper](https://arxiv.org/abs/2108.08291) | [code](https://github.com/cvg/pixel-perfect-sfm)<br><br>
@@ -1132,6 +1213,9 @@
 
 ### 场景图生成(Scene Graph Generation)
 
+[5] Graph-to-3D: End-to-End Generation and Manipulation of 3D Scenes Using Scene Graphs<br>
+[paper](https://arxiv.org/abs/2108.08841)<br><br>
+
 [4] Target Adaptive Context Aggregation for Video Scene Graph Generation<br>
 [paper](https://arxiv.org/abs/2108.08121) | [code](https://github.com/MCG-NJU/TRACE)<br><br>
 
@@ -1162,6 +1246,9 @@
 
 ### 数据增广(Data Augmentation)
 
+[2] Amplitude-Phase Recombination: Rethinking Robustness of Convolutional Neural Networks in Frequency Domain<br>
+[paper](https://arxiv.org/abs/2108.08487) | [code](https://github/iCGY96/APR)<br><br>
+
 [1] MixMo: Mixing Multiple Inputs for Multiple Outputs via Deep Subnetworks<br>
 [paper](https://arxiv.org/abs/2103.06132)<br>
 [解读：“白嫖”性能的MixMo，一种新的数据增强or模型融合方法](https://zhuanlan.zhihu.com/p/396554361)<br><br>
@@ -1203,6 +1290,9 @@
 
 ### 图像聚类(Image Clustering)
 
+[5] A Unified Objective for Novel Class Discovery(Oral)<br>
+[paper](https://arxiv.org/abs/2108.08536) | [code](https://ncd-uno.github.io/)<br><br>
+
 [4] Instance Similarity Learning for Unsupervised Feature Representation<br>
 [paper](https://arxiv.org/abs/2108.02721) | [code](https://github.com/ZiweiWangTHU/ISL.git)<br><br>
 
@@ -1330,6 +1420,9 @@
 
 ## 对比学习(Contrastive Learning)
 
+[5] Self-Supervised Video Representation Learning with Meta-Contrastive Network(对比学习)(元学习)(表征学习)(动作识别)<br>
+[paper](https://arxiv.org/abs/2108.08426)<br><br>
+
 [4] Improving Contrastive Learning by Visualizing Feature Transformation<br>
 [paper](https://arxiv.org/abs/2108.02982) ｜ [visualization tools and codes](https://github.com/DTennant/CL-Visualizing-Feature-Transformation)<br><br>
 
@@ -1373,6 +1466,9 @@
 
 ## 元学习(Meta Learning)
 
+[1] Self-Supervised Video Representation Learning with Meta-Contrastive Network(对比学习)(元学习)(表征学习)(动作识别)<br>
+[paper](https://arxiv.org/abs/2108.08426)<br><br>
+
 <br>
 <a name="MMLearning"/> 
 
@@ -1392,6 +1488,9 @@
 
 ## 视觉预测(Vision-based Prediction)
 
+[8] Generating Smooth Pose Sequences for Diverse Human Motion Prediction<br>
+[paper](https://arxiv.org/abs/2108.08422) | [code](https://github.com/wei-mao-2019/gsps)<br><br>
+
 [7] MSR-GCN: Multi-Scale Residual Graph Convolution Networks for Human Motion Prediction(人体运动预测)<br>
 [paper](https://arxiv.org/abs/2108.07152) ｜ [code](https://github.com/Droliven/MSRGCN)<br><br>
 
@@ -1516,6 +1615,9 @@ Mixed SIGNals: Sign Language Production via a Mixture of Motion Primitives(手
 Temporal-wise Attention Spiking Neural Networks for Event Streams Classification<br>
 [paper](https://arxiv.org/abs/2107.11711)<br><br>
 
+Click to Move: Controlling Video Generation with Sparse Motion<br>
+[paper](https://arxiv.org/abs/2108.08815) | [code](https://github.com/PierfrancescoArdino/C2M)<br><br>
+
 Long-Term Temporally Consistent Unpaired Video Translation from Simulated Surgical 3D Data（视频翻译/医学/视频合成）<br>
 [paper](https://arxiv.org/abs/2103.17204)<br><br>
 
@@ -1537,6 +1639,12 @@ KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs<br>
 
 # 2. ICCV2021 Oral（更新中）
 
+[30] DECA: Deep viewpoint-Equivariant human pose estimation using Capsule Autoencoders(Oral)<br>
+[paper](https://arxiv.org/abs/2108.08557) | [code](https://github.com/mmlab-cv/DECA)<br><br>
+
+[29] A Unified Objective for Novel Class Discovery(Oral)<br>
+[paper](https://arxiv.org/abs/2108.08536) | [code](https://ncd-uno.github.io/)<br><br>
+
 [28] Multi-Anchor Active Domain Adaptation for Semantic Segmentation(Oral)<br>
 [paper](https://arxiv.org/abs/2108.08012)<br><br>