Monocular depth estimation with SPN loss

作者：

Highlights：

• We estimate direct depth instead of disparity unlike previous work.

• Novel loss function to take advantage of the dissimilarity of projected images.

• We used camera geometric projection with stereo cameras unlike previous work.

• Achieved state of the art depth accuracy on KITTI test dataset.

摘要

•We estimate direct depth instead of disparity unlike previous work.•Novel loss function to take advantage of the dissimilarity of projected images.•We used camera geometric projection with stereo cameras unlike previous work.•Achieved state of the art depth accuracy on KITTI test dataset.

论文关键词：Depth estimation,Monocular depth estimation

论文评审过程：Received 10 August 2019, Revised 24 February 2020, Accepted 12 May 2020, Available online 19 May 2020, Version of Record 21 May 2020.

论文官网地址：https://doi.org/10.1016/j.imavis.2020.103934