Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981575
Cicero L. Costa, Túlia A. A. Macedo, C. Barcelos
Pelvic dysfunction mainly affects adult women, it is estimated that 15% of multiparous women suffer from the problem. Dysfunctions can be diagnosed by defecography, a dynamic MRI scan. Images are used by specialists to diagnose organ dysfunction such as the bladder and the early rectum. This paper presents an automated classification system that uses a non-rigid registration based on a variational model to create automatic markings from initial markings made by an expert. The classification is based on simple average and the centroids of the K-means grouping technique. The classification made by the system is evaluated by confusion matrix based metrics. The obtained results using 21 defecography exams from 21 different patients indicate that the proposed technique is a promising tool in the diagnosis of pelvic floor disorders and can assist the physician in the diagnostic process.
{"title":"Pre-diagnosis of pelvic floor disorders-based image registration and clustering","authors":"Cicero L. Costa, Túlia A. A. Macedo, C. Barcelos","doi":"10.1109/ICAR46387.2019.8981575","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981575","url":null,"abstract":"Pelvic dysfunction mainly affects adult women, it is estimated that 15% of multiparous women suffer from the problem. Dysfunctions can be diagnosed by defecography, a dynamic MRI scan. Images are used by specialists to diagnose organ dysfunction such as the bladder and the early rectum. This paper presents an automated classification system that uses a non-rigid registration based on a variational model to create automatic markings from initial markings made by an expert. The classification is based on simple average and the centroids of the K-means grouping technique. The classification made by the system is evaluated by confusion matrix based metrics. The obtained results using 21 defecography exams from 21 different patients indicate that the proposed technique is a promising tool in the diagnosis of pelvic floor disorders and can assist the physician in the diagnostic process.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"25 1","pages":"572-577"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82230441","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981632
Amy Deeb, M. Seto, Yajun Pan
Navigation in dynamic environments is a challenge for autonomous vehicles operating without prior maps or global position references. This poses high risk to vehicles that perform scientific studies and monitoring missions in marine Arctic environments characterized by slowly moving sea ice with few truly static landmarks. Whereas mature simultaneous localization and mapping (SLAM) approaches assume a static environment, this work extends pose graph SLAM to spatiotemporally evolving environments. A novel model-based dynamic factor is proposed to capture a landmark's state transition model - whether the state be kinematic, appearance or otherwise. The structure of the state transition model is assumed to be known a priori, while the parameters are estimated on-line. Expectation maximization is used to avoid adding variables to the graph. Proof-of-concept results are shown in small- and medium-scale simulation, and small-scale laboratory environments for a small quadrotor. Preliminary laboratory validation results shows the effect of mechanical limitations of the quadrotor platform and increased uncertainties associated with the model-based dynamic factors on the SLAM estimate. Simulation results are encouraging for the application of model-based dynamic factors to dynamic landmarks with a constant-velocity kinematic model.
{"title":"Model-based Dynamic Pose Graph SLAM in Unstructured Dynamic Environments","authors":"Amy Deeb, M. Seto, Yajun Pan","doi":"10.1109/ICAR46387.2019.8981632","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981632","url":null,"abstract":"Navigation in dynamic environments is a challenge for autonomous vehicles operating without prior maps or global position references. This poses high risk to vehicles that perform scientific studies and monitoring missions in marine Arctic environments characterized by slowly moving sea ice with few truly static landmarks. Whereas mature simultaneous localization and mapping (SLAM) approaches assume a static environment, this work extends pose graph SLAM to spatiotemporally evolving environments. A novel model-based dynamic factor is proposed to capture a landmark's state transition model - whether the state be kinematic, appearance or otherwise. The structure of the state transition model is assumed to be known a priori, while the parameters are estimated on-line. Expectation maximization is used to avoid adding variables to the graph. Proof-of-concept results are shown in small- and medium-scale simulation, and small-scale laboratory environments for a small quadrotor. Preliminary laboratory validation results shows the effect of mechanical limitations of the quadrotor platform and increased uncertainties associated with the model-based dynamic factors on the SLAM estimate. Simulation results are encouraging for the application of model-based dynamic factors to dynamic landmarks with a constant-velocity kinematic model.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"78 1","pages":"123-128"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84967194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981627
Héctor Azpúrua, Filipe A. S. Rocha, G. Garcia, Alexandre Souza Santos, Eduardo Cota, Luiz Guilherme Barros, Alexandre S Thiago, G. Pessin, G. Freitas
Robotic devices operating in confined environments such as underground, tunnel systems, and cave networks are currently receiving particular attention from research and industry. One example is the Brazilian mining company Vale S.A., which is employing a robot - Espeleolcobô - to access restricted areas. The robot was designed initially to inspect natural caves during teleoperated missions and is now being used to monitor dam galleries and other confined or dangerous environments. This paper describes recent developments regarding locomotion mechanisms and mobility analyses, localization, and path planning strategies aiming autonomous robot operation. Preliminary results from simulation and field experiments validate the robotic device concept. Lessons learned from multiple field tests in industrial scenarios are also described.
{"title":"EspeleoRobô - a robotic device to inspect confined environments","authors":"Héctor Azpúrua, Filipe A. S. Rocha, G. Garcia, Alexandre Souza Santos, Eduardo Cota, Luiz Guilherme Barros, Alexandre S Thiago, G. Pessin, G. Freitas","doi":"10.1109/ICAR46387.2019.8981627","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981627","url":null,"abstract":"Robotic devices operating in confined environments such as underground, tunnel systems, and cave networks are currently receiving particular attention from research and industry. One example is the Brazilian mining company Vale S.A., which is employing a robot - Espeleolcobô - to access restricted areas. The robot was designed initially to inspect natural caves during teleoperated missions and is now being used to monitor dam galleries and other confined or dangerous environments. This paper describes recent developments regarding locomotion mechanisms and mobility analyses, localization, and path planning strategies aiming autonomous robot operation. Preliminary results from simulation and field experiments validate the robotic device concept. Lessons learned from multiple field tests in industrial scenarios are also described.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"75 1","pages":"17-23"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83380279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981653
Edison Kleiber Titito Concha, Diego Pittol, Ricardo Westhauser, M. Kolberg, R. Maffei, Edson Prestes e Silva
Keyframe-based monocular SLAM (Simultaneous Localization and Mapping) is one of the main visual SLAM approaches, used to estimate the camera motion together with the map reconstruction over selected frames. These techniques represent the environment by map points located in the three-dimensional space, that can be recognized and located in the frame. However, these techniques usually cannot decide when a map point is an outlier or obsolete information and can be discarded. Another problem is to decide when combining map points corresponding to the same three-dimensional point. In this paper, we present a robust method to maintain a refined map. This approach uses the covisibility graph and an algorithm based on information fusion to build a probabilistic map, that explicitly models outlier measurements. In addition, we incorporate a pruning mechanism to reduce redundant information and remove outliers. In this way, our approach manages to reduce the map size maintaining essential information of the environment. Finally, in order to evaluate the performance of our method, we incorporate it into an ORB-SLAM system and measure the accuracy achieved on publicly available benchmark datasets which contain indoor images sequences recorded with a hand-held monocular camera.
基于关键帧的单眼SLAM (Simultaneous Localization and Mapping)是主要的视觉SLAM方法之一，用于估计摄像机运动并在选定帧上重建地图。这些技术通过位于三维空间中的地图点来表示环境，这些点可以在框架中被识别和定位。然而，这些技术通常不能确定一个地图点何时是一个异常值或过时的信息，可以丢弃。另一个问题是决定何时组合对应于同一三维点的地图点。在本文中，我们提出了一种鲁棒的方法来维护一个精细化的映射。该方法使用共可见度图和基于信息融合的算法来构建概率图，明确地对离群值进行建模。此外，我们还采用了修剪机制来减少冗余信息和去除异常值。通过这种方式，我们的方法设法减小了地图尺寸，同时保持了环境的基本信息。最后，为了评估我们的方法的性能，我们将其纳入ORB-SLAM系统，并测量了在公开可用的基准数据集上实现的精度，这些基准数据集包含用手持单目相机记录的室内图像序列。
{"title":"Map Point Optimization in Keyframe-Based SLAM using Covisibility Graph and Information Fusion","authors":"Edison Kleiber Titito Concha, Diego Pittol, Ricardo Westhauser, M. Kolberg, R. Maffei, Edson Prestes e Silva","doi":"10.1109/ICAR46387.2019.8981653","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981653","url":null,"abstract":"Keyframe-based monocular SLAM (Simultaneous Localization and Mapping) is one of the main visual SLAM approaches, used to estimate the camera motion together with the map reconstruction over selected frames. These techniques represent the environment by map points located in the three-dimensional space, that can be recognized and located in the frame. However, these techniques usually cannot decide when a map point is an outlier or obsolete information and can be discarded. Another problem is to decide when combining map points corresponding to the same three-dimensional point. In this paper, we present a robust method to maintain a refined map. This approach uses the covisibility graph and an algorithm based on information fusion to build a probabilistic map, that explicitly models outlier measurements. In addition, we incorporate a pruning mechanism to reduce redundant information and remove outliers. In this way, our approach manages to reduce the map size maintaining essential information of the environment. Finally, in order to evaluate the performance of our method, we incorporate it into an ORB-SLAM system and measure the accuracy achieved on publicly available benchmark datasets which contain indoor images sequences recorded with a hand-held monocular camera.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"24 1","pages":"129-134"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88606538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981651
E. G. Ribeiro, V. Grassi
The development of the robotics field has not yet allowed robots to execute, with dexterity, simple actions performed by humans. One of them is the grasping of objects by robotic manipulators. Aiming to explore the use of deep learning algorithms, specifically Convolutional Neural Networks (CNN), to approach the robotic grasping problem, this work addresses the visual perception phase involved in the task. To achieve this goal, the dataset “Cornell Grasp” was used to train a CNN capable of predicting the most suitable place to grasp the object. It does this by obtaining a grasping rectangle that symbolizes the position, orientation, and opening of the robot's parallel grippers just before the grippers are closed. The proposed system works in real-time due to the small number of network parameters. This is possible by means of the data augmentation strategy used. The efficiency of the detection is in accordance with the state of the art and the speed of prediction, to the best of our knowledge, is the highest in the literature.
{"title":"Fast Convolutional Neural Network for Real-Time Robotic Grasp Detection","authors":"E. G. Ribeiro, V. Grassi","doi":"10.1109/ICAR46387.2019.8981651","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981651","url":null,"abstract":"The development of the robotics field has not yet allowed robots to execute, with dexterity, simple actions performed by humans. One of them is the grasping of objects by robotic manipulators. Aiming to explore the use of deep learning algorithms, specifically Convolutional Neural Networks (CNN), to approach the robotic grasping problem, this work addresses the visual perception phase involved in the task. To achieve this goal, the dataset “Cornell Grasp” was used to train a CNN capable of predicting the most suitable place to grasp the object. It does this by obtaining a grasping rectangle that symbolizes the position, orientation, and opening of the robot's parallel grippers just before the grippers are closed. The proposed system works in real-time due to the small number of network parameters. This is possible by means of the data augmentation strategy used. The efficiency of the detection is in accordance with the state of the art and the speed of prediction, to the best of our knowledge, is the highest in the literature.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"129 1","pages":"49-54"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75266096","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981636
J. C. Vendrichoski, T. L. Costa, E. S. Elyoussef, E. D. Pieri
With a wide spectrum of applications ranging from entertainment to military use, quadrotors, in their conventional construction with four fixed rotors, have proven to be sufficiently capable of robustly performing numerous tasks. However, the capability to generate thrust just in one direction, i.e., normal to its main plane, is very restrictive and drastically reduces the maneuverability and agility of the vehicle. In this paper, an alternative model of the quadrotor is presented. Constructively, the difference lies in the addition of a mechanism that tilts the rotors in the longitudinal direction, which in practice adds maneuverability by enabling the longitudinal translation uncoupled from the pitching movement. In addition to this new thrust component in the longitudinal direction, this configuration also yields a significant increase in the yaw torque. These are highly desired features in a UAV used to execute tasks that require physical interaction with the surrounding environment. The mathematical model of the entire system is obtained by employing the Euler-Lagrange formalism and a multi-body approach. In addition, a basic control scheme is used to verify, through simulation, the obtained model.
{"title":"Mathematical modeling and control of a quadrotor aerial vehicle with tiltrotors aimed for interaction tasks","authors":"J. C. Vendrichoski, T. L. Costa, E. S. Elyoussef, E. D. Pieri","doi":"10.1109/ICAR46387.2019.8981636","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981636","url":null,"abstract":"With a wide spectrum of applications ranging from entertainment to military use, quadrotors, in their conventional construction with four fixed rotors, have proven to be sufficiently capable of robustly performing numerous tasks. However, the capability to generate thrust just in one direction, i.e., normal to its main plane, is very restrictive and drastically reduces the maneuverability and agility of the vehicle. In this paper, an alternative model of the quadrotor is presented. Constructively, the difference lies in the addition of a mechanism that tilts the rotors in the longitudinal direction, which in practice adds maneuverability by enabling the longitudinal translation uncoupled from the pitching movement. In addition to this new thrust component in the longitudinal direction, this configuration also yields a significant increase in the yaw torque. These are highly desired features in a UAV used to execute tasks that require physical interaction with the surrounding environment. The mathematical model of the entire system is obtained by employing the Euler-Lagrange formalism and a multi-body approach. In addition, a basic control scheme is used to verify, through simulation, the obtained model.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"38 1","pages":"161-166"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79388832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981628
N. Dhanaraj, Julietta Maffeo, G. Pereira, Jason N. Gross, Yu Gu, Nathan Hewitt, Casey Edmonds-Estes, Rachel Jarman, Jeongwoo Seo, Henry Gunner, Alexandra Hatfield, Tucker Johnson, Lunet Yifru
This paper describes the Adaptable Platform for Interactive Swarm robotics (APIS) - a testbed designed to accelerate development in human-swarm interaction (HSI) research. Specifically, this paper presents the design of a swarm robot platform composed of fifty low cost robots coupled with a testing field and a software architecture that allows for modular and versatile development of swarm algorithms. The motivation behind developing this platform is that the emergence of a swarm's collective behavior can be difficult to predict and control. However, human-swarm interaction can measurably increase a swarm's performance as the human operator may have intuition or knowledge unavailable to the swarm. The development of APIS allows researchers to focus on HSI research, without being constrained to a fixed ruleset or interface. A short survey is presented that offers a taxonomy of swarm platforms and provides conclusions that contextualize the development of APIS. Next, the motivations, design and functionality of the APIS testbed are described. Finally, the operation and potential of the platform are demonstrated through two experimental evaluations.
{"title":"Adaptable Platform for Interactive Swarm Robotics (APIS): A Human-Swarm Interaction Research Testbed","authors":"N. Dhanaraj, Julietta Maffeo, G. Pereira, Jason N. Gross, Yu Gu, Nathan Hewitt, Casey Edmonds-Estes, Rachel Jarman, Jeongwoo Seo, Henry Gunner, Alexandra Hatfield, Tucker Johnson, Lunet Yifru","doi":"10.1109/ICAR46387.2019.8981628","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981628","url":null,"abstract":"This paper describes the Adaptable Platform for Interactive Swarm robotics (APIS) - a testbed designed to accelerate development in human-swarm interaction (HSI) research. Specifically, this paper presents the design of a swarm robot platform composed of fifty low cost robots coupled with a testing field and a software architecture that allows for modular and versatile development of swarm algorithms. The motivation behind developing this platform is that the emergence of a swarm's collective behavior can be difficult to predict and control. However, human-swarm interaction can measurably increase a swarm's performance as the human operator may have intuition or knowledge unavailable to the swarm. The development of APIS allows researchers to focus on HSI research, without being constrained to a fixed ruleset or interface. A short survey is presented that offers a taxonomy of swarm platforms and provides conclusions that contextualize the development of APIS. Next, the motivations, design and functionality of the APIS testbed are described. Finally, the operation and potential of the platform are demonstrated through two experimental evaluations.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"31 1","pages":"720-726"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75015681","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981570
Kyle Lindgren, Sarah Leung, W. Nothwang, E. J. Shamwell
Machine learning has emerged as an extraordinary tool for solving many computer vision tasks by extracting and correlating meaningful features from high dimensional inputs in ways that often exceed the best human-derived modeling efforts. However, the area of vision-aided localization remains diverse with many traditional, model-based approaches (i.e. filtering- or nonlinear least- squares- based) often outperforming deep, model-free approaches. In this work, we present Bootstrapped Monocular VIO (BooM), a scaled monocular visual-inertial odometry (VIO) solution that leverages the complex data association ability of model-free approaches with the ability to exploit known geometric dynamics with model-based approaches. Our end-to-end, unsupervised deep neural network simultaneously learns to perform visual-inertial odometry and estimate scene depth while scale is enforced through a loss signal computed from position change magnitude estimates from traditional methods. We evaluate our network against a state-of-the-art (SoA) approach on the KITTI driving dataset as well as a micro aerial vehicle (MAV) dataset that we collected in the AirSim simulation environment. We further demonstrate the benefits of our combined approach through robustness tests on degraded trajectories.
{"title":"BooM-Vio: Bootstrapped Monocular Visual-Inertial Odometry with Absolute Trajectory Estimation through Unsupervised Deep Learning","authors":"Kyle Lindgren, Sarah Leung, W. Nothwang, E. J. Shamwell","doi":"10.1109/ICAR46387.2019.8981570","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981570","url":null,"abstract":"Machine learning has emerged as an extraordinary tool for solving many computer vision tasks by extracting and correlating meaningful features from high dimensional inputs in ways that often exceed the best human-derived modeling efforts. However, the area of vision-aided localization remains diverse with many traditional, model-based approaches (i.e. filtering- or nonlinear least- squares- based) often outperforming deep, model-free approaches. In this work, we present Bootstrapped Monocular VIO (BooM), a scaled monocular visual-inertial odometry (VIO) solution that leverages the complex data association ability of model-free approaches with the ability to exploit known geometric dynamics with model-based approaches. Our end-to-end, unsupervised deep neural network simultaneously learns to perform visual-inertial odometry and estimate scene depth while scale is enforced through a loss signal computed from position change magnitude estimates from traditional methods. We evaluate our network against a state-of-the-art (SoA) approach on the KITTI driving dataset as well as a micro aerial vehicle (MAV) dataset that we collected in the AirSim simulation environment. We further demonstrate the benefits of our combined approach through robustness tests on degraded trajectories.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"71 1","pages":"516-522"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89708563","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981633
Yuchi Ishikawa, Haruya Ishikawa, S. Akizuki, Masaki Yamazaki, Y. Taniguchi, Y. Aoki
We propose novel representations for functions of an object, namely Task-oriented Function, which is improved upon the idea of Afforadance in the field of Robotics Vision. We also propose a convolutional neural network to detect task-oriented functions. This network takes as input an operational task as well as an RGB image and assign each pixel an appropriate label for every task. Task-oriented funciton makes it possible to descibe various ways to use an object because the outputs from the network differ depending on operational tasks. We introduce a new dataset for task-oriented function detection, which contains about 1200 RGB images and 6000 pixel-level annotations assuming five tasks. Our proposed method reached 0.80 mean IOU in our dataset.
{"title":"Task-oriented Function Detection Based on Operational Tasks","authors":"Yuchi Ishikawa, Haruya Ishikawa, S. Akizuki, Masaki Yamazaki, Y. Taniguchi, Y. Aoki","doi":"10.1109/ICAR46387.2019.8981633","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981633","url":null,"abstract":"We propose novel representations for functions of an object, namely Task-oriented Function, which is improved upon the idea of Afforadance in the field of Robotics Vision. We also propose a convolutional neural network to detect task-oriented functions. This network takes as input an operational task as well as an RGB image and assign each pixel an appropriate label for every task. Task-oriented funciton makes it possible to descibe various ways to use an object because the outputs from the network differ depending on operational tasks. We introduce a new dataset for task-oriented function detection, which contains about 1200 RGB images and 6000 pixel-level annotations assuming five tasks. Our proposed method reached 0.80 mean IOU in our dataset.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"36 1","pages":"635-640"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89736537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/ICAR46387.2019.8981553
J. Smet, E. V. Poorten, V. Poliakov, Kenan Niu, Frédérique Chesterman, J. Fornier, M. Ahmad, M. Ourak, Viktor Vörös, J. Deprest
During laparoscopic sacrocolpopexy, pelvic organ prolapse is repaired by suturing one side of a synthetic mesh around the vaginal vault while stapling the other end to the sacrum, restoring the anatomical position of the vagina. A perineal assistant positions and tensions the vault with a vaginal manipulator instrument to properly expose the vaginal tissue to the laparoscopic surgeon. A technical difficulty during this surgery is the loss of depth perception due to visualization of the patient's internals on a 2D screen. Especially during precise surgical tasks, a more natural way to understand the distance between the laparoscopic instruments and the surgical region of interest could be advantageous. This work describes an exploratory study to investigate the potential of introducing 3D visualization into this surgical intervention. More in particular, experimentation is conducted with autostereoscopic display technology. A mixed reality setup was constructed featuring a virtual reality model of the vagina, 2D and 3D visualization, a physical interface representing the tissue of the body wall and a tracking system to track instrument motion. An experiment was conducted whereby the participants had to navigate the instrument to a number of pre-defined locations under 2D or 3D visualization. Compared to 2D, a considerable reduction in average task time (-42.9 %), travelled path lenght (-31.8 %) and errors (-52.2 %) was observed when performing the experiment in 3D. Where this work demonstrated a potential benefit of autostereoscopic visualization with respect to 2D visualization, in future work we wish to investigate if there also exists a benefit when comparing this technology with conventional stereoscopic visualization and whether stereoscopy can be used for (semi-) automated guidance during robotic laparoscopy.
{"title":"Evaluating the Potential Benefit of Autostereoscopy in Laparoscopic Sacrocolpopexy through VR Simulation","authors":"J. Smet, E. V. Poorten, V. Poliakov, Kenan Niu, Frédérique Chesterman, J. Fornier, M. Ahmad, M. Ourak, Viktor Vörös, J. Deprest","doi":"10.1109/ICAR46387.2019.8981553","DOIUrl":"https://doi.org/10.1109/ICAR46387.2019.8981553","url":null,"abstract":"During laparoscopic sacrocolpopexy, pelvic organ prolapse is repaired by suturing one side of a synthetic mesh around the vaginal vault while stapling the other end to the sacrum, restoring the anatomical position of the vagina. A perineal assistant positions and tensions the vault with a vaginal manipulator instrument to properly expose the vaginal tissue to the laparoscopic surgeon. A technical difficulty during this surgery is the loss of depth perception due to visualization of the patient's internals on a 2D screen. Especially during precise surgical tasks, a more natural way to understand the distance between the laparoscopic instruments and the surgical region of interest could be advantageous. This work describes an exploratory study to investigate the potential of introducing 3D visualization into this surgical intervention. More in particular, experimentation is conducted with autostereoscopic display technology. A mixed reality setup was constructed featuring a virtual reality model of the vagina, 2D and 3D visualization, a physical interface representing the tissue of the body wall and a tracking system to track instrument motion. An experiment was conducted whereby the participants had to navigate the instrument to a number of pre-defined locations under 2D or 3D visualization. Compared to 2D, a considerable reduction in average task time (-42.9 %), travelled path lenght (-31.8 %) and errors (-52.2 %) was observed when performing the experiment in 3D. Where this work demonstrated a potential benefit of autostereoscopic visualization with respect to 2D visualization, in future work we wish to investigate if there also exists a benefit when comparing this technology with conventional stereoscopic visualization and whether stereoscopy can be used for (semi-) automated guidance during robotic laparoscopy.","PeriodicalId":6606,"journal":{"name":"2019 19th International Conference on Advanced Robotics (ICAR)","volume":"26 1","pages":"566-571"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91029830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：