定量结果显示,SC-DepthV1取得了与Monodepth2相持平的结果。但SC-DepthV1估计出的深度图具有连续性,因此还是SC-DepthV1更胜一筹。 值得一提的是,SC-DepthV1在论文里除了介绍自己的定量结果外,还将单目深度估计结果引入到了ORB-SLAM2中,构建了一个伪RGB-D SLAM系统。在一些序列中,Monodepth2+ORB-SLAM2的系统...
SC-Depth (V1, V2, and V3) for Unsupervised Monocular Depth Estimation Webpage:https://jiawangbian.github.io/sc_depth_pl/ - cudnn/sc_depth_pl
It is integrated into SC-DepthV1 and jointly trained with self-supervised losses, greatly boosting the performance. In the SC-DepthV3 (ArXiv 2022), we propose a robust learning framework for accurate and sharp monocular depth estimation in (highly) dynamic scenes. As the photometric loss, ...
PoE 1 PC servers are restarting in: 00:51:49 . They should be back up in approximately10 minutes. PoE 1 Xbox servers are restarting in: 00:51:48 . They should be back up in approximately10 minutes. PoE 1 Sony servers are restarting in: ...
unsupervised monocular depth estimation in indoor scenes. Based on this findings, we propose auto-recitify network (ARN) to handle the large relative rotation between consecutive video frames. It is integrated into SC-DepthV1 and jointly trained with self-supervised losses, greatly boosting the ...
python eval_depth.py \ --dataset$DATASET_FOLDER\ --pred_depth=$RESULTS_FOLDER\ --gt_depth=$GT_FOLDER\ --seg_mask=$SEG_MASK_FOLDER References SC-DepthV1: Unsupervised Scale-consistent Depth Learning from Video (IJCV 2021) Jia-Wang Bian, Huangying Zhan, Naiyan Wang, Zhichao Li, Le Zhang...
unsupervised monocular depth estimation in indoor scenes. Based on this findings, we propose auto-recitify network (ARN) to handle the large relative rotation between consecutive video frames. It is integrated into SC-DepthV1 and jointly trained with self-supervised losses, greatly boosting the ...
unsupervised monocular depth estimation in indoor scenes. Based on this findings, we propose auto-recitify network (ARN) to handle the large relative rotation between consecutive video frames. It is integrated into SC-DepthV1 and jointly trained with self-supervised losses, greatly boosting the ...
unsupervised monocular depth estimation in indoor scenes. Based on this findings, we propose auto-recitify network (ARN) to handle the large relative rotation between consecutive video frames. It is integrated into SC-DepthV1 and jointly trained with self-supervised losses, greatly boosting the ...
python eval_depth.py \ --dataset$DATASET_FOLDER\ --pred_depth=$RESULTS_FOLDER\ --gt_depth=$GT_FOLDER\ --seg_mask=$SEG_MASK_FOLDER References SC-DepthV1: Unsupervised Scale-consistent Depth Learning from Video (IJCV 2021) Jia-Wang Bian, Huangying Zhan, Naiyan Wang, Zhichao Li, Le Zhang...