CV

• 深度估计
• AP解释https://www.zhihu.com/question/53405779
• yolov4中的route和shortcut层
• 双线性插值https://blog.csdn.net/qq_14845119/article/details/107557449
• Focal Loss
• 相机模型、参数和各个坐标系
• 目标检测中的AP,mAP
• Nerf
  基于query获得结果的思想 Nerf和Transformer/Detr差异
  ECCV 2020 NeRF的 paper, (Mildenhall, Ben, et al. "Nerf: Representing scenes as neural radiance fields for view synthesis." ECCV, 2020.)
• Diffusion模型
  https://zhuanlan.zhihu.com/p/493533589
  https://www.bilibili.com/video/BV1Je4y127nH/?spm_id_from=444.41.list.card_archive.click&vd_source=e61b4912071bd625bd7a88745457bf40
• BEV Nerf中，Nerf预测出来的网格点x,y的高度z,一起组成3D坐标点，使用高度渲染投影回2D图片，这里的高度渲染怎么做的？ 为什么高度渲染不可微？ 渲染可以使用世界坐标系也可以使用局部坐标系，都怎么实现的？
• 端到端自车位姿校正，论文调研：
  CNN位姿估计
  understanding the limitations of cnn-based absolute camera pose regression
  CoordiNet: uncertainty-aware pose regressor for reliable vehicle localization
  Geometry-Aware Learning of Maps for Camera Localization
  光流估计：
  PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume
  Raft: Recurrent all-pairs field transforms for optical flow
• 矢量建图
• 视觉的目的
• Nerf:
  
• Diffusion Models
  
• vision-language model训练与应用
  
• https://cvpr2022.thecvf.com/workshop-schedule
  Workshop on Autonomous Driving (WAD)
  The 1st Workshop on Vision Datasets Understanding
  Visual Perception and Learning in an Open World
  What can computer vision learn from visual neuroscience?
  5th MUltimodal Learning and Applications Workshop (MULA)
  Image Matching: Local Features and Beyond
  Efficient Deep Learning for Computer Vision
  Workshop on Attention and Transformers in Vision
• 2022 CVPR WAD
  MOTS
  https://cvpr2022.wad.vision/
• 3D OD
  https://zhuanlan.zhihu.com/p/432135656
• https://lilianweng.github.io/