The hexadecimal color code #34526f is a medium dark shade of cyan-blue. The human body masks, derived from the segmentation model, are. Features include: ; Automatic lecture scheduling and access management coupled with CAMPUSOnline ; Livestreaming from lecture halls ; Support for Extron SMPs and automatic backup. 4. de belongs to TUM-RBG, DE. Bei Fragen steht unser Helpdesk gerne zur Verfügung! RBG Helpdesk. 73% improvements in high-dynamic scenarios. Last update: 2021/02/04. We select images in dynamic scenes for testing. t. The data was recorded at full frame rate (30 Hz) and sensor resolution (640x480). TUM dataset contains the RGB and Depth images of Microsoft Kinect sensor along the ground-truth trajectory of the sensor. tum-rbg (RIPE) Prefix status Active, Allocated under RIPE Size of prefixThe TUM RGB-D benchmark for visual odometry and SLAM evaluation is presented and the evaluation results of the first users from outside the group are discussed and briefly summarized. Current 3D edge points are projected into reference frames. tum. We select images in dynamic scenes for testing. First, both depths are related by a deformation that depends on the image content. Attention: This is a live. libs contains options for training, testing and custom dataloaders for TUM, NYU, KITTI datasets. Previously, I worked on fusing RGB-D data into 3D scene representations in real-time and improving the quality of such reconstructions with various deep learning approaches. We increased the localization accuracy and mapping effects compared with two state-of-the-art object SLAM algorithms. It contains indoor sequences from RGB-D sensors grouped in several categories by different texture, illumination and structure conditions. We also provide a ROS node to process live monocular, stereo or RGB-D streams. tum. de / [email protected]. 0 is a lightweight and easy-to-set-up Windows tool that works great for Gigabyte and non-Gigabyte users who’re just starting out with RGB synchronization. Compared with art-of-the-state methods, experiments on the TUM RBG-D dataset, KITTI odometry dataset, and practical environment show that SVG-Loop has advantages in complex environments with varying light, changeable weather, and. navab}@tum. ASN details for every IP address and every ASN’s related domains, allocation date, registry name, total number of IP addresses, and assigned prefixes. General Info Open in Search Geo: Germany (DE) — Domain: tum. Only RGB images in sequences were applied to verify different methods. This paper presents a novel unsupervised framework for estimating single-view depth and predicting camera motion jointly. Seen 143 times between April 1st, 2023 and April 1st, 2023. vmcarle35. We also provide a ROS node to process live monocular, stereo or RGB-D streams. tum. It is a significant component in V-SLAM (Visual Simultaneous Localization and Mapping) systems. the initializer is very slow, and does not work very reliably. de. tum. Both groups of sequences have important challenges such as missing depth data caused by sensor. tum. Living room has 3D surface ground truth together with the depth-maps as well as camera poses and as a result perfectly suits not just for benchmarking camera. The TUM RGB-D dataset, published by TUM Computer Vision Group in 2012, consists of 39 sequences recorded at 30 frames per second using a Microsoft Kinect sensor in different indoor scenes. 2. 2-pack RGB lights can fill light in multi-direction. Welcome to the RBG user central. de. Deep learning has promoted the. de TUM-Live. de) or your attending physician can advise you in this regard. TUM dataset contains the RGB and Depth images of Microsoft Kinect sensor along the ground-truth trajectory of the sensor. Telephone: 089 289 18018. 4. dePrinting via the web in Qpilot. Monocular SLAM PTAM [18] is a monocular, keyframe-based SLAM system which was the first work to introduce the idea of splitting camera tracking and mapping into parallel threads, and. Exercises will be held remotely and live on the Thursday slot about each 3 to 4 weeks and will not be recorded. The RGB-D case shows the keyframe poses estimated in sequence fr1 room from the TUM RGB-D Dataset [3], andNote. We provide examples to run the SLAM system in the KITTI dataset as stereo or monocular, in the TUM dataset as RGB-D or monocular, and in the EuRoC dataset as stereo or monocular. [SUN RGB-D] The SUN RGB-D dataset contains 10,335 RGBD images with semantic labels organized in 37. rbg. Major Features include a modern UI with dark-mode Support and a Live-Chat. net. General Info Open in Search Geo: Germany (DE) — Domain: tum. We provide a large dataset containing RGB-D data and ground-truth data with the goal to establish a novel benchmark for the evaluation of visual odometry and visual SLAM systems. Visual SLAM (VSLAM) has been developing rapidly due to its advantages of low-cost sensors, the easy fusion of other sensors, and richer environmental information. Tickets: [email protected]. These tasks are being resolved by one Simultaneous Localization and Mapping module called SLAM. TUM RGB-Dand RGB-D inputs. TUM RGB-D dataset. We evaluated ReFusion on the TUM RGB-D dataset [17], as well as on our own dataset, showing the versatility and robustness of our approach, reaching in several scenes equal or better performance than other dense SLAM approaches. Mystic Light. in. Qualified applicants please apply online at the link below. In order to obtain the missing depth information of the pixels in current frame, a frame-constrained depth-fusion approach has been developed using the past frames in a local window. rbg. The RGB-D dataset contains the following. The experiment on the TUM RGB-D dataset shows that the system can operate stably in a highly dynamic environment and significantly improve the accuracy of the camera trajectory. Our method named DP-SLAM is implemented on the public TUM RGB-D dataset. Our methodTUM-Live, the livestreaming and VoD service of the Rechnerbetriebsgruppe at the department of informatics and mathematics at the Technical University of Munichon RGB-D data. rbg. These sequences are separated into two categories: low-dynamic scenarios and high-dynamic scenarios. tum- / RBG-account is entirely seperate form the LRZ- / TUM-credentials. Mystic Light. Bauer Hörsaal (5602. Do you know your RBG. A challenging problem in SLAM is the inferior tracking performance in the low-texture environment due to their low-level feature based tactic. while in the challenging TUM RGB-D dataset, we use 30 iterations for tracking, with max keyframe interval µ k = 5. There are two persons sitting at a desk. Evaluation of Localization and Mapping Evaluation on Replica. Compared with art-of-the-state methods, experiments on the TUM RBG-D dataset, KITTI odometry dataset, and practical environment show that SVG-Loop has advantages in complex environments with varying light, changeable weather, and dynamic interference. - GitHub - raulmur/evaluate_ate_scale: Modified tool of the TUM RGB-D dataset that automatically computes the optimal scale factor that aligns trajectory and groundtruth. The single and multi-view fusion we propose is challenging in several aspects. 89. Network 131. TUM RBG-D can be used with TUM RGB-D or UZH trajectory evaluation tools and has the following format timestamp[s] tx ty tz qx qy qz qw. Usage. sh","path":"_download. X. Google Scholar: Access. TUM data set contains three sequences, in which fr1 and fr2 are static scene data sets, and fr3 is dynamic scene data sets. It is able to detect loops and relocalize the camera in real time. 7 nm. Registrar: RIPENCC Route. SLAM with Standard Datasets KITTI Odometry dataset . RBG. de which are continuously updated. usage: generate_pointcloud. We recorded a large set of image sequences from a Microsoft Kinect with highly accurate and time-synchronized ground truth camera poses from a motion capture system. This is in contrast to public SLAM benchmarks like e. TUM RGB-D Benchmark Dataset [11] is a large dataset containing RGB-D data and ground-truth camera poses. : You need VPN ( VPN Chair) to open the Qpilot Website. Features include: Automatic lecture scheduling and access management coupled with CAMPUSOnline. It is able to detect loops and relocalize the camera in real time. Moreover, the metric. however, the code for the orichid color is E6A8D7, not C0448F as it says, since it already belongs to red violet. Map Points: A list of 3-D points that represent the map of the environment reconstructed from the key frames. We exclude the scenes with NaN poses generated by BundleFusion. Maybe replace by your own way to get an initialization. An Open3D RGBDImage is composed of two images, RGBDImage. The depth images are already registered w. In this work, we add the RGB-L (LiDAR) mode to the well-known ORB-SLAM3. The number of RGB-D images is 154, each with a corresponding scribble and a ground truth image. Open3D has a data structure for images. 1. However, the pose estimation accuracy of ORB-SLAM2 degrades when a significant part of the scene is occupied by moving ob-jects (e. Our experimental results have showed the proposed SLAM system outperforms the ORB. Similar behaviour is observed in other vSLAM [23] and VO [12] systems as well. Experimental results on the TUM RGB-D dataset and our own sequences demonstrate that our approach can improve performance of state-of-the-art SLAM system in various challenging scenarios. It lists all image files in the dataset. 德国慕尼黑工业大学TUM计算机视觉组2012年提出了一个RGB-D数据集，是目前应用最为广泛的RGB-D数据集。数据集使用Kinect采集，包含了depth图像和rgb图像，以及ground truth等数据，具体格式请查看官网。on the TUM RGB-D dataset. TUM RGB-D Scribble-based Segmentation Benchmark Description. de (The registered domain) AS: AS209335 - TUM-RBG, DE Note: An IP might be announced by multiple ASs. TE-ORB_SLAM2 is a work that investigate two different methods to improve the tracking of ORB-SLAM2 in. RGB-live. de from your own Computer via Secure Shell. The data was recorded at full frame rate (30 Hz) and sensor res-olution 640 480. The TUM RGB-D dataset , which includes 39 sequences of offices, was selected as the indoor dataset to test the SVG-Loop algorithm. Once this works, you might want to try the 'desk' dataset, which covers four tables and contains several loop closures. VPN-Connection to the TUM set up of the RBG certificate Furthermore the helpdesk maintains two websites. In this paper, we present the TUM RGB-D benchmark for visual odometry and SLAM evaluation and report on the first use-cases and users of it outside our own group. Login (with in. de TUM-RBG, DE. Further details can be found in the related publication. unicorn. The categorization differentiates. Example result (left are without dynamic object detection or masks, right are with YOLOv3 and masks), run on rgbd_dataset_freiburg3_walking_xyz: Getting Started. In this paper, we present the TUM RGB-D bench-mark for visual odometry and SLAM evaluation and report on the first use-cases and users of it outside our own group. Our method named DP-SLAM is implemented on the public TUM RGB-D dataset. The results indicate that the proposed DT-SLAM (mean RMSE = 0:0807. TUM-Live, the livestreaming and VoD service of the Rechnerbetriebsgruppe at the department of informatics and mathematics at the Technical University of Munichand RGB-D inputs. de. The key constituent of simultaneous localization and mapping (SLAM) is the joint optimization of sensor trajectory estimation and 3D map construction. RGB and HEX color codes of TUM colors. in. Visual Odometry is an important area of information fusion in which the central aim is to estimate the pose of a robot using data collected by visual sensors. Then, the unstable feature points are removed, thus. I received my MSc in Informatics in the summer of 2019 at TUM and before that, my BSc in Informatics and Multimedia at the University of Augsburg. The dataset has RGB-D sequences with ground truth camera trajectories. Loop closure detection is an important component of Simultaneous. The fr1 and fr2 sequences of the dataset are employed in the experiments, which contain scenes of a middle-sized office and an industrial hall environment respectively. 04 on a computer (i7-9700K CPU, 16 GB RAM and Nvidia GeForce RTX 2060 GPU). Many answers for common questions can be found quickly in those articles. A robot equipped with a vision sensor uses the visual data provided by cameras to estimate the position and orientation of the robot with respect to its surroundings [11]. Change your RBG-Credentials. 89 papers with code • 0 benchmarks • 20 datasets. Welcome to the Introduction to Deep Learning course offered in SS22. github","path":". Tracking: Once a map is initialized, the pose of the camera is estimated for each new RGB-D image by matching features in. An Open3D Image can be directly converted to/from a numpy array. Semantic navigation based on the object-level map, a more robust. In the following section of this paper, we provide the framework of the proposed method OC-SLAM with the modules in the semantic object detection thread and dense mapping thread. We provide a large dataset containing RGB-D data and ground-truth data with the goal to establish a novel benchmark for the evaluation of visual odometry and visual SLAM systems. The TUM RGB-D dataset consists of colour and depth images (640 × 480) acquired by a Microsoft Kinect sensor at a full frame rate (30 Hz). 2. The experiments on the public TUM dataset show that, compared with ORB-SLAM2, the MOR-SLAM improves the absolute trajectory accuracy by 95. The following seven sequences used in this analysis depict different situations and intended to test robustness of algorithms in these conditions. Cookies help us deliver our services. Simultaneous localization and mapping (SLAM) is one of the fundamental capabilities for intelligent mobile robots to perform state estimation in unknown environments. 1 Linux and Mac OS; 1. Lecture 1: Introduction Tuesday, 10/18/2022, 05:00 AM. Modified tool of the TUM RGB-D dataset that automatically computes the optimal scale factor that aligns trajectory and groundtruth. Welcome to the RBG-Helpdesk! What kind of assistance do we offer? The Rechnerbetriebsgruppe (RBG) maintaines the infrastructure of the Faculties of Computer. The actions can be generally divided into three categories: 40 daily actions (e. The sequences contain both the color and depth images in full sensor resolution (640 × 480). October. IROS, 2012. net. tum. Thus, there will be a live stream and the recording will be provided. If you want to contribute, please create a pull request and just wait for it to be reviewed ;) An RGB-D camera is commonly used for mobile robots, which is low-cost and commercially available. in. It contains indoor sequences from RGB-D sensors grouped in several categories by different texture, illumination and structure conditions. globalAuf dieser Seite findet sich alles Wissenwerte zum guten Start mit den Diensten der RBG. 8%(except Completion Ratio) improvement in accuracy compared to NICE-SLAM [14]. The results indicate that the proposed DT-SLAM (mean RMSE = 0:0807. Follow us on: News. 22 Dec 2016: Added AR demo (see section 7). Example result (left are without dynamic object detection or masks, right are with YOLOv3 and masks), run on rgbd_dataset_freiburg3_walking_xyz: Getting Started. The system is also integrated with Robot Operating System (ROS) [10], and its performance is verified by testing DS-SLAM on a robot in a real environment. Visual odometry and SLAM datasets: The TUM RGB-D dataset [14] is focused on the evaluation of RGB-D odometry and SLAM algorithms and has been extensively used by the research community. We provided an. Not observed on urlscan. Simultaneous Localization and Mapping is now widely adopted by many applications, and researchers have produced very dense literature on this topic. tum. and TUM RGB-D [42], our framework is shown to outperform both monocular SLAM system (i. , drinking, eating, reading), nine health-related actions (e. 1. 4. SUNCG is a large-scale dataset of synthetic 3D scenes with dense volumetric annotations. tum. In these situations, traditional VSLAMInvalid Request. The measurement of the depth images is millimeter. de / [email protected](PTR record of primary IP) Recent Screenshots. dataset [35] and real-world TUM RGB-D dataset [32] are two benchmarks widely used to compare and analyze 3D scene reconstruction systems in terms of camera pose estimation and surface reconstruction. Two different scenes (the living room and the office room scene) are provided with ground truth. 1 Performance evaluation on TUM RGB-D dataset The TUM RGB-D dataset was proposed by the TUM Computer Vision Group in 2012, which is frequently used in the SLAM domain [ 6 ]. 0/16 (Route of ASN) PTR: unicorn. Hotline: 089/289-18018. 1. Estimating the camera trajectory from an RGB-D image stream: TODO. We also provide a ROS node to process live monocular, stereo or RGB-D streams. The Wiki wiki. Only RGB images in sequences were applied to verify different methods. in. from publication: DDL-SLAM: A robust RGB-D SLAM in dynamic environments combined with Deep. , 2012). As an accurate pose tracking technique for dynamic environments, our efficient approach utilizing CRF-based long-term consistency can estimate a camera trajectory (red) close to the ground truth (green). Registrar: RIPENCC Recent Screenshots. 0. de Performance evaluation on TUM RGB-D dataset This study uses the Freiburg3 series from the TUM RGB-D dataset. Montiel and Dorian Galvez-Lopez 13 Jan 2017: OpenCV 3 and Eigen 3. This dataset was collected by a Kinect V1 camera at the Technical University of Munich in 2012. via a shortcut or the back-button); Cookies are. tum. This repository is the collection of SLAM-related datasets. The proposed V-SLAM has been tested on public TUM RGB-D dataset. /build/run_tum_rgbd_slam Allowed options: -h, --help produce help message -v, --vocab arg vocabulary file path -d, --data-dir arg directory path which contains dataset -c, --config arg config file path --frame-skip arg (=1) interval of frame skip --no-sleep not wait for next frame in real time --auto-term automatically terminate the viewer --debug. 159. employees/guests and hiwis have an ITO account and the print account has been added to the ITO account. Our dataset contains the color and depth images of a Microsoft Kinect sensor along the ground-truth trajectory of the sensor. However, the method of handling outliers in actual data directly affects the accuracy of. de(PTR record of primary IP) IPv4: 131. Standard ViT Architecture . de which are continuously updated. /Datasets/Demo folder. 5. in. Awesome visual place recognition (VPR) datasets. See the list of other web pages hosted by TUM-RBG, DE. 230A tag already exists with the provided branch name. Deep learning has promoted the. The calibration of the RGB camera is the following: fx = 542. 73 and 2a09:80c0:2::73 . The results indicate that the proposed DT-SLAM (mean RMSE = 0:0807. Digitally Addressable RGB (DRGB) allows you to color each LED individually, rather than choosing one static color for the entire LED strip, meaning you can go full rainbow. Ultimately, Section. 0/16 (Route of ASN) Recent Screenshots. 822841 fy = 542. A pose graph is a graph in which the nodes represent pose estimates and are connected by edges representing the relative poses between nodes with measurement uncertainty [23]. The Wiki wiki. system is evaluated on TUM RGB-D dataset [9]. {"payload":{"allShortcutsEnabled":false,"fileTree":{"":{"items":[{"name":". TUM rgb-d data set contains rgb-d image. We tested the proposed SLAM system on the popular TUM RGB-D benchmark dataset . We provide examples to run the SLAM system in the KITTI dataset as stereo or monocular, in the TUM dataset as RGB-D or monocular, and in the EuRoC dataset as stereo or monocular. We also provide a ROS node to process live monocular, stereo or RGB-D streams. tum. Thumbnail Figures from Complex Urban, NCLT, Oxford robotcar, KiTTi, Cityscapes datasets. We provide examples to run the SLAM system in the KITTI dataset as stereo or monocular, in the TUM dataset as RGB-D or monocular, and in the EuRoC dataset as stereo or monocular. tum. tum. We recommend that you use the 'xyz' series for your first experiments. Object–object association. 289. 576870 cx = 315. 2. +49. This is not shown. RBG – Rechnerbetriebsgruppe Mathematik und Informatik Helpdesk: Montag bis Freitag 08:00 - 18:00 Uhr Telefon: 18018 Mail: rbg@in. In [19], the authors tested and analyzed the performance of selected visual odometry algorithms designed for RGB-D sensors on the TUM dataset with respect to accuracy, time, and memory consumption. The datasets we picked for evaluation are listed below and the results are summarized in Table 1. Dependencies: requirements. 2023. This is contributed by the fact that the maximum consensus out-Compared with art-of-the-state methods, experiments on the TUM RBG-D dataset, KITTI odometry dataset, and practical environment show that SVG-Loop has advantages in complex environments with varying light, changeable weather, and. rbg. 0. Export as Portable Document Format (PDF) using the Web BrowserExport as PDF, XML, TEX or BIB. 0. de / rbg@ma. 02. It includes 39 indoor scene sequences, of which we selected dynamic sequences to evaluate our system. The dynamic objects have been segmented and removed in these synthetic images. To stimulate comparison, we propose two evaluation metrics and provide automatic evaluation tools. The RGB-D dataset[3] has been popular in SLAM research and was a benchmark for comparison too. Traditional visual SLAM algorithms run robustly under the assumption of a static environment, but always fail in dynamic scenarios, since moving objects will impair camera pose tracking. 0. RBG – Rechnerbetriebsgruppe Mathematik und Informatik Helpdesk: Montag bis Freitag 08:00 - 18:00 Uhr Telefon: 18018 Mail: rbg@in. C. RGB-D cameras that can provide rich 2D visual and 3D depth information are well suited to the motion estimation of indoor mobile robots. Digitally Addressable RGB. tum. Experimental results show , the combined SLAM system can construct a semantic octree map with more complete and stable semantic information in dynamic scenes. Contribution. For interference caused by indoor moving objects, we add the improved lightweight object detection network YOLOv4-tiny to detect dynamic regions, and the dynamic features in the dynamic area are then eliminated in the algorithm. This paper presents a novel SLAM system which leverages feature-wise. Account activation. e. © RBG Rechnerbetriebsgruppe Informatik, Technische Universität München, 2013–2018, rbg@in. tum. Last update: 2021/02/04. deAwesome SLAM Datasets. The button save_traj saves the trajectory in one of two formats (euroc_fmt or tum_rgbd_fmt). tum. This zone conveys a joint 2D and 3D information corresponding to the distance of a given pixel to the nearest human body and the depth distance to the nearest human, respectively. It contains walking, sitting and desk sequences, and the walking sequences are mainly utilized for our experiments, since they are highly dynamic scenarios where two persons are walking back and forth. Freiburg3 consists of a high-dynamic scene sequence marked 'walking', in which two people walk around a table, and a low-dynamic scene sequence marked 'sitting', in which two people sit in chairs with slight head or part. Direct. The benchmark website contains the dataset, evaluation tools and additional information. With the advent of smart devices, embedding cameras, inertial measurement units, visual SLAM (vSLAM), and visual-inertial SLAM (viSLAM) are enabling novel general public. No incoming hits Nothing talked to this IP. The desk sequence describes a scene in which a person sits. We extensively evaluate the system on the widely used TUM RGB-D dataset, which contains sequences of small to large-scale indoor environments, with respect to different parameter combinations. objects—scheme [6]. tum. 1 freiburg2 desk with person The TUM dataset is a well-known dataset for evaluating SLAM systems in indoor environments. In particular, RGB ORB-SLAM fails on walking_xyz, while pRGBD-Refined succeeds and achieves the best performance on. de credentials) Kontakt Rechnerbetriebsgruppe der Fakultäten Mathematik und Informatik Telefon: 18018 rbg@in. Large-scale experiments are conducted on the ScanNet dataset, showing that volumetric methods with our geometry integration mechanism outperform state-of-the-art methods quantitatively as well as qualitatively. Experimental results on the TUM RGB-D and the KITTI stereo datasets demonstrate our superiority over the state-of-the-art. rbg. 1. The sequences are from TUM RGB-D dataset. Invite others by sharing the room link and access code. IEEE/RJS International Conference on Intelligent Robot, 2012. Each light has 260 LED beads and high CRI 95+, which makes the pictures and videos taken more natural and beautiful. Qualitative and quantitative experiments show that our method outperforms state-of-the-art approaches in various dynamic scenes in terms of both accuracy and robustness. sequences of some dynamic scenes, and has the accurate. The ICL-NUIM dataset aims at benchmarking RGB-D, Visual Odometry and SLAM algorithms. github","contentType":"directory"},{"name":". vehicles) [31]. WePDF. The session will take place on Monday, 25. The standard training and test set contain 795 and 654 images, respectively. Motchallenge. The stereo case shows the final trajectory and sparse reconstruction of the sequence 00 from the KITTI dataset [2]. 0/16. 0/16 Abuse Contact data. This project was created to redesign the Livestream and VoD website of the RBG-Multimedia group. In addition, results on real-world TUM RGB-D dataset also gain agreement with the previous work (Klose, Heise, and Knoll Citation 2013) in which IC can slightly increase the convergence radius and improve the precision in some sequences (e. in. Information Technology Technical University of Munich Arcisstr. Contribution. This repository is the collection of SLAM-related datasets. de (The registered domain) AS: AS209335 - TUM-RBG, DE Note: An IP might be announced by multiple ASs. [2] She was nominated by President Bill Clinton to replace retiring justice. Here, RGB-D refers to a dataset with both RGB (color) images and Depth images. (TUM) RGB-D data set show that the presented scheme outperforms the state-of-art RGB-D SLAM systems in terms of trajectory. Livestream on Artemis → Lectures or live. de(PTR record of primary IP) IPv4: 131. Single-view depth captures the local structure of mid-level regions, including texture-less areas, but the estimated depth lacks global coherence. two example RGB frames from a dynamic scene and the resulting model built by our approach. 16% green and 43. KITTI Odometry dataset is a benchmarking dataset for monocular and stereo visual odometry and lidar odometry that is captured from car-mounted devices. There are multiple configuration variants: standard - general purpose; 2. Technische Universität München, TU München, TUM), заснований в 1868 році, знаходиться в місті Мюнхені і є єдиним технічним університетом Баварії і одним з найбільших вищих навчальних закладів у. io. Registrar: RIPENCC Route: 131. While previous datasets were used for object recognition, this dataset is used to understand the geometry of a scene. 159. New College Dataset. However, these DATMO. Our extensive experiments on three standard datasets, Replica, ScanNet, and TUM RGB-D show that ESLAM improves the accuracy of 3D reconstruction and camera localization of state-of-the-art dense visual SLAM methods by more than 50%, while it runs up to 10 times faster and does not require any pre-training.