整理：AI算法与图像处理
CVPR2022论文和代码整理：https://github.com/DWCTOD/CVPR2022-Papers-with-Code-Demo
ECCV2022论文和代码整理：https://github.com/DWCTOD/ECCV2022-Papers-with-Code-Demo
最新成果demo展示：
标题：Teaching Where to Look: Attention Similarity Knowledge Distillation for Low Resolution Face Recognition
论文：https://arxiv.org/abs/2209.14498
代码：https://github.com/gist-ailab/teaching-where-to-look
深度学习在人脸识别基准测试中取得了优异的性能，但在低分辨率（LR）图像中性能显著下降。我们提出了一种注意力相似性知识提取方法，该方法将作为教师从高分辨率（HR）网络获得的注意力图转换为作为学生的LR网络，以提高LR识别性能。受人类能够基于HR图像中获得的先验知识从LR图像中近似物体区域的启发，我们利用余弦相似性设计了知识提取损失，使学生网络的注意力与教师网络的注意力相似。在各种LR人脸相关基准上的实验证实，所提出的方法总体上改善了LR设置下的识别性能，通过简单地传输构造良好的注意力图，优于最新的结果
最新论文整理
ECCV2022
Updated on : 25 Oct 2022
total number : 5
Learning Neural Radiance Fields from Multi-View Geometry
- 论文/Paper: http://arxiv.org/pdf/2210.13041
- 代码/Code: None
UIA-ViT: Unsupervised Inconsistency-Aware Method based on Vision Transformer for Face Forgery Detection
- 论文/Paper: http://arxiv.org/pdf/2210.12752
- 代码/Code: None
SC-wLS: Towards Interpretable Feed-forward Camera Re-localization
- 论文/Paper: http://arxiv.org/pdf/2210.12748
- 代码/Code: https://github.com/xinwu98/sc-wls
Photo-realistic Neural Domain Randomization
- 论文/Paper: http://arxiv.org/pdf/2210.12682
- 代码/Code: None
HAM: Hierarchical Attention Model with High Performance for 3D Visual Grounding
- 论文/Paper: http://arxiv.org/pdf/2210.12513
- 代码/Code: None
CVPR2022
NeurIPS
Updated on : 25 Oct 2022
total number : 16
Monocular Dynamic View Synthesis: A Reality Check
- 论文/Paper: http://arxiv.org/pdf/2210.13445
- 代码/Code: https://github.com/kair-bair/dycheck
Contrastive Representation Learning for Gaze Estimation
- 论文/Paper: http://arxiv.org/pdf/2210.13404
- 代码/Code: https://github.com/jswati31/gazeclr
High-Resolution Image Editing via Multi-Stage Blended Diffusion
- 论文/Paper: http://arxiv.org/pdf/2210.12965
- 代码/Code: None
Revisiting Sparse Convolutional Model for Visual Recognition
- 论文/Paper: http://arxiv.org/pdf/2210.12945
- 代码/Code: https://github.com/Delay-Xili/SDNet.
360-MLC: Multi-view Layout Consistency for Self-training and Hyper-parameter Tuning
- 论文/Paper: http://arxiv.org/pdf/2210.12935
- 代码/Code: None
Respecting Transfer Gap in Knowledge Distillation
- 论文/Paper: http://arxiv.org/pdf/2210.12787
- 代码/Code: None
Beta R-CNN: Looking into Pedestrian Detection from Another Perspective
- 论文/Paper: http://arxiv.org/pdf/2210.12758
- 代码/Code: None
Cut-and-Approximate: 3D Shape Reconstruction from Planar Cross-sections with Deep Reinforcement Learning
- 论文/Paper: http://arxiv.org/pdf/2210.12509
- 代码/Code: None
NeuPhysics: Editable Neural Geometry and Physics from Monocular Videos
- 论文/Paper: http://arxiv.org/pdf/2210.12352
- 代码/Code: None
Attention-Based Scattering Network for Satellite Imagery
- 论文/Paper: http://arxiv.org/pdf/2210.12185
- 代码/Code: None
NVIDIA FLARE: Federated Learning from Simulation to Real-World
- 论文/Paper: http://arxiv.org/pdf/2210.13291
- 代码/Code: https://github.com/nvidia/nvflare
Deep Equilibrium Approaches to Diffusion Models
- 论文/Paper: http://arxiv.org/pdf/2210.12867
- 代码/Code: https://github.com/locuslab/deq-ddim
Functional Indirection Neural Estimator for Better Out-of-distribution Generalization
- 论文/Paper: http://arxiv.org/pdf/2210.12739
- 代码/Code: None
Outsourcing Training without Uploading Data via Efficient Collaborative Open-Source Sampling
- 论文/Paper: http://arxiv.org/pdf/2210.12575
- 代码/Code: None
Generative Modeling of High-resolution Global Precipitation Forecasts
- 论文/Paper: http://arxiv.org/pdf/2210.12504
- 代码/Code: None
Score-based Denoising Diffusion with Non-Isotropic Gaussian Noise Models
- 论文/Paper: http://arxiv.org/pdf/2210.12254
- 代码/Code: None
