graviti
产品服务
解决方案
知识库
公开数据集
关于我们
SBM-RGBD Dataset
Autonomous Driving
|...
许可协议: Unknown

Overview

The SBM-RGBD dataset [provides] all facilities (data,ground truths, and evaluation scripts) in order to evaluateand compare scene background modelling methods formoving object detection on RGBD videos. It includes35 RGBD videos acquired by the Microsoft Kinect andprovided as synchronised colour and depth sequences. Theseare representative of typical indoor visual data captured invideo surveillance and smart environment scenarios [...].The videos span 7 categories, selected to include diversescene background modelling challenges for moving objectdetection related only to the RGB channels, only to thedepth channel, or related to all the channels: IlluminationChanges, Color Camouflage, Depth Camouflage,Intermittent Motion, Out of Sensor Range, Shadows, andBootstrapping.To enable a precise quantitative comparison and rankingof various algorithms for moving object detection fromRGBD videos, each video comes with a set of pixel-wiseground truth foreground segmentations. Moreover, thedataset comes with tools to compute performance metricsfor moving object detection from RGBD videos, and thusidentify algorithms that are robust across various challenges.The SBM-RGBD dataset has been created for the SBMRGBDChallenge, organized in conjunction with theWorkshop on Background Learning for Detection andTracking from RGBD Videos (RGBD2017), but it willremain available, together with the Challenge results, alsoafter the competition, as reference for future methods.M. Camplani, L. Maddalena, G. MoyàAlcover, A. Petrosino, L. Salgado, A BenchmarkingFramework for Background Subtraction in RGBD videos,in S. Battiato, G. Gallo, G.M. Farinella, M. Leo (Eds),New Trends in Image Analysis and Processing-ICIAP2017 Workshops, Lecture Notes in Computer Science,Springer, 2017

数据概要
数据格式
video, image,
数据量
--
文件大小
--
| 数据量 -- | 大小 --
SBM-RGBD Dataset
Autonomous Driving
许可协议: Unknown

Overview

The SBM-RGBD dataset [provides] all facilities (data,ground truths, and evaluation scripts) in order to evaluateand compare scene background modelling methods formoving object detection on RGBD videos. It includes35 RGBD videos acquired by the Microsoft Kinect andprovided as synchronised colour and depth sequences. Theseare representative of typical indoor visual data captured invideo surveillance and smart environment scenarios [...].The videos span 7 categories, selected to include diversescene background modelling challenges for moving objectdetection related only to the RGB channels, only to thedepth channel, or related to all the channels: IlluminationChanges, Color Camouflage, Depth Camouflage,Intermittent Motion, Out of Sensor Range, Shadows, andBootstrapping.To enable a precise quantitative comparison and rankingof various algorithms for moving object detection fromRGBD videos, each video comes with a set of pixel-wiseground truth foreground segmentations. Moreover, thedataset comes with tools to compute performance metricsfor moving object detection from RGBD videos, and thusidentify algorithms that are robust across various challenges.The SBM-RGBD dataset has been created for the SBMRGBDChallenge, organized in conjunction with theWorkshop on Background Learning for Detection andTracking from RGBD Videos (RGBD2017), but it willremain available, together with the Challenge results, alsoafter the competition, as reference for future methods.M. Camplani, L. Maddalena, G. MoyàAlcover, A. Petrosino, L. Salgado, A BenchmarkingFramework for Background Subtraction in RGBD videos,in S. Battiato, G. Gallo, G.M. Farinella, M. Leo (Eds),New Trends in Image Analysis and Processing-ICIAP2017 Workshops, Lecture Notes in Computer Science,Springer, 2017

0
立即开始构建AI
graviti
wechat-QR
长按保存识别二维码,关注Graviti公众号

Copyright@Graviti
沪ICP备19019574号
沪公网安备 31011002004865号