← All publications
Propagating Confidences through CNNs for Sparse Data Regression
Abstract
Introduces normalized convolutional layers that propagate per-pixel confidence alongside sparse input data, a foundational technique for uncertainty-aware sparse-data regression.