学习日记2020-12-7论文学习

Multi-Scale Progressive Fusion Learning for Depth Map Super-Resolution

1.depth encoder branch, reconstruction branch， color encoder branch.
2.LR Model and HR Model
3.loss function

HFS: Hierarchical Feature Selection for Efficient Image Segmentation

1.a careful system implementation on modern GPUs for efficient feature computation;
2.) an effective hierarchical feature selection and fusion strategy with learning.

Struck: Structured Output T racking with Kernel

1.erform additional experiments to quantify the effects of structured learning
2.be implemented on the GPU and demonstrate how high frame-rates can be achieved without sacrificing tracking performance
？？？？

DenseCut: Densely Connected CRFs for Realtime GrabCut

SemanticPaint: Interactive 3D Labeling and Learning at your Fingertips