# camera pose/depth paper list
# depth review article
- Monocular depth estimation based on deep learning: An overview(2020)
- Deep learning for monocular depth estimation: A review(2021)
# perspective camera
## supervised
- Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer
- 用很多不同的dataset 包含stereo movie
- depth model for single image
- 提出針對 dataset 間 depth 尺度不同的解法 (算 loss 時先校正 depth scale)
## Unsupervised
### depth+ego motion
- Unsupervised Learning of Depth and Ego-Motion from Video (CVPR 2017)
- with phtotometric loss
- start of unsupervised method
- https://github.com/tinghuiz/SfMLearner
- Jointly train pose net and depth net
- Depth from Videos in the Wild: Unsupervised Monocular Depth Learning from Unknown Cameras
- Joint train pose net and depth net
- Occlusion awarness & forground/ background recognition
- Unsupervised Scale-consistent Depth Learning from Video(2021)
- introducing geometric loss (the one Cindy's stablization paper used)
- this geometric loss will sometimes fail and regress to homogeneous depthmap
- https://github.com/JiawangBian/SC-SfMLearner-Release
- Monodepth2: Digging Into Self-Supervised Monocular Depth Estimation
- 之前Cindy報過
- Per-Pixel Minimum Reprojection Loss
- delt with eocclusion
- Auto-Masking Stationary Pixels
### hybrid method
- Consistent Video Depth Estimation
- use geometric loss to finetune sfm path
- Robust Consistent Video Depth Estimation
- built on "Consistent Video Depth Estimation"
- Pose and Depth with Flexible Deformation(the transformation can be different between each pixels between frames)
- blurry depth
- depth filter
- bring back detail of depth

### Visual Odometry
- D3VO: Deep Depth, Deep Pose and Deep Uncertainty for Monocular Visual Odometry
- DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras
# Temporal consistency
- Enforcing Temporal Consistency in Video Depth Estimation (ICCV 2021)

# 360 camera
## mapping of 360 info
- Sphere mapping for feature extraction from 360° fish-eye captures (我覺得這好像滿重要ㄉ)
- 360 Depth Estimation in the Wild - the Depth360 Dataset and the SegFuse Network (VR 2022)
- 用網路上360影片 optical flow, camera pose 做的 depth，也有給 Net

# Tutorial
- Introduction to Camera Pose Estimation with Deep Learning (CVPR 2019)
- https://arxiv.org/pdf/1907.05272.pdf