Deep hierarchical guidance and regularization learning for end-to-end depth estimation

作者：

Highlights：

• We propose a Hierarchical Guidance and Regularization (HGR) learning framework for end-to-end monocular depth estimation.

• A multi-regularized learning strategy is to optimize network parameters by employing multi-level information of depth maps.

• The proposed method obtains state-of-the-art depth estimation performance on NYU Depth V2, KITTI and Make3D datasets.

摘要

•We propose a Hierarchical Guidance and Regularization (HGR) learning framework for end-to-end monocular depth estimation.•A multi-regularized learning strategy is to optimize network parameters by employing multi-level information of depth maps.•The proposed method obtains state-of-the-art depth estimation performance on NYU Depth V2, KITTI and Make3D datasets.

论文关键词：Depth estimation,Multi-regularization,Deep neural network

论文评审过程：Received 3 March 2017, Revised 1 May 2018, Accepted 13 May 2018, Available online 23 June 2018, Version of Record 23 June 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.05.016