SFU Yasutaka Furukawa
3D Reconstruction From Sensing, Perception, to Cognition
sensing->perception->cognition
Tango, CANVAS by Occipital
sensing revolution
sensor data
Calibration and camera poses
Geometric model
SfM, SLAM, Matching-moving
Visual Octometry
Multi-view
A Theory of Shape by Space Carving (Algorithm)
Plane Sweep Algorithm
Pixel -> Patch
Global Optimization
Multi-view High consistency
matching, feature matching (depth)
Markov Random Field
概率模型
图模型
三角不等式
-Expansion
Tree-reweighted Message Passing
Maximum Peak vs MRF optimization
View Clustering
low level -> high level vs high level -> low level
CSG reconstruction
floorplan
RGBD Video Stream
Corner Candidate
PlaneRCNN
（英语太差了）