Through extensive evaluations, we show that 3D dis- tillation significantly improves the depth accuracy of reflective surfaces on ScanNet [4] and 7-Scenes [39],.
Improving Self-Supervised Monocular Depth Estimation on Reflective ...
ieeexplore.ieee.org › document
In our experiments using the ScanNet and 7-Scenes datasets, we show that 3D distillation not only significantly improves the prediction accuracy, especially on ...
We propose 3D distillation: a novel training framework that utilizes the projected depth of reconstructed reflective surfaces to generate reasonably accurate ...
To improve the depth prediction accuracy for reflective surfaces without increasing the computational cost, Shi et al. [49] utilized the pseudo labels, obtained ...
Sep 27, 2024 · TL;DR: We propose a reflective-aware triplet loss and multi-teacher knowledge distillation, improving depth accuracy on reflective surfaces.
3D Distillation: Improving Self-Supervised Monocular Depth Estimation on Reflective Surfaces ; Issue Date: 2023-10-04 ; Language: English.
3D Distillation: Improving Self-Supervised Monocular Depth Estimation on Reflective Surfaces, ICCV 2023; Mirror3D: Depth Refinement for Mirror Surfaces, CVPR ...
Co-authors ; 3D Distillation: Improving Self-Supervised Monocular Depth Estimation on Reflective Surfaces. X Shi, G Dikov, G Reitmayr, TK Kim, M Ghafoorian.
Self-supervised monocular depth estimation (SSMDE) aims at predicting the dense depth maps of monocular images, by learning to minimize a photometric loss using ...
Nov 21, 2023 · In the next ICCV paper “3D Distillation: Improving Self-Supervised Monocular Depth Estimation on Reflective Surfaces,” we propose a new method.