Hariharan, Bharath, et al. “Hypercolumns for object segmentation and fine-grained localization.” Proceedings of the IEEE Conference on Computer Vision and Pattern Recogni-
tion. 2015. (Citations: 185).
tion. 2015. (Citations: 185).
1 Motivation
The high layer features care more about “what” but lose localization information, while the low layer features care more about “where” but are not category-level sensitive enough. See Fig.
2 Pipeline
See Fig. This is used in the refinement step. Upsampling is used to resize feature maps to have same size.