scripts/download 加入第一个模型,SC-Depth-V1 Dec 12, 2024 system 修复验证阶段提示,添加MidAir配置及总配置文件 Jan 2, 2025 tools 修复验证阶段提示,添加MidAir配置及总配置文件 Jan 2, 2025 utils 修复验证阶段提示,添加MidAir配置及总配置文件 Jan 2, 2025 ...
SC-Depth (V1, V2, and V3) for Unsupervised Monocular Depth Estimation Webpage:https://jiawangbian.github.io/sc_depth_pl/ - JiawangBian/sc_depth_pl
作者:Jia-Wang Bian, Huangying Zhan, Naiyan Wang, Tat-Jun Chin, Chunhua Shen, Ian Reid代码链接:https://github.com/JiawangBian/sc_depth_pl文章链接:https://arxiv.org/abs/2006.02708v2 一、研究背景 无监督单目深度估计算法【1】仅使用无需标注的单目视频作为训练数据就能在测试时提供较为准确的场景...
项目链接:https://github.com/JiawangBian/sc_depth_pl Talk·嘉宾介绍 边佳旺牛津大学博士后研究员 边佳旺,牛津大学博士后研究员。博士毕业于澳大利亚阿德莱德大学,本科毕业于南开大学计算机系。研究方向为三维计算机视觉,具体包括图像特征匹配,单目深度估计,和基于NeRF的三维重建和渲染等。边佳旺在CVPR,NeurIPS,IJCV, TPA...
我们提出了一个单目深度估计器SCDepth,它只需要无标记的视频进行训练,并能在推断时进行尺度一致的预测。我们的贡献包括:(i)我们提出了几何一致性损失,这将惩罚相邻视图之间的预测深度不一致;(ii)我们提出了一个自我发现的掩码来自动定位那些在训练过程中违反静态场景假设并引起噪声信号的运动目标;(iii)我们通过详细的...
作者: Jia-Wang Bian, Huangying Zhan, Naiyan Wang, Tat-Jun Chin, Chunhua Shen, Ian Reid 代码链接: https://github.com/JiawangBian/sc_depth_pl 文章链接: https://arxiv.org/abs/2006.02708v2 一、研究背景 无监督单目深度估计算法【1】仅使用无需标注的单目视频作为训练数据就能在测试时提供较为准确的...
Watchers 0 watching Forks 0 forks Report repository Releases No releases published Packages No packages published Languages Python 98.1% Shell 1.9% Footer © 2025 GitHub, Inc. Footer navigation Terms Privacy Security Status Docs Contact Manage cookies Do not share my personal information ...
Pleaser refer to our new implementation of SC-Depth (V1, V2, and V3) at https://github.com/JiawangBian/sc_depth_pl This codebase implements the SC-DepthV1 described in the paper: Unsupervised Scale-consistent Depth Learning from Video Jia-Wang Bian, Huangying Zhan, Naiyan Wang, Zhichao ...
import numpy as np import torch from kornia.geometry.depth import depth_to_normals from pytorch_lightning import LightningModule import losses.loss_functions as LossF from models.DepthNet import DepthNet from models.PoseNet import PoseNet from visualization import * class SC_DepthV3(LightningModule):...
This is the offical codes for the methods described in the "Feature-metric Loss for Self-supervised Learning of Depth and Egomotion". - sconlyshootery/FeatDepth