Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 23
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Very fast solution to the PnP problem with algebraic outlier rejection

  • Ferraz, Luis
  • Binefa, Xavier
  • Moreno-Noguer, Francesc
Presentado al CVPR 2014 celebrado en Columbus, Ohio (US) del 23 al 28 de junio., We propose a real-time, robust to outliers and accurate solution to the Perspective-n-Point (PnP) problem. The main advantages of our solution are twofold: first, it integrates the outlier rejection within the pose estimation pipeline with a negligible computational overhead; and second, its scalability to arbitrarily large number of correspondences. Given a set of 3D-to-2D matches, we formulate pose estimation problem as a low-rank homogeneous system where the solution lies on its 1D null space. Outlier correspondences are those rows of the linear system which perturb the null space and are progressively detected by projecting them on an iteratively estimated solution of the null space. Since our outlier removal process is based on an algebraic criterion which does not require computing the full-pose and reprojecting back all 3D points on the image plane at each step, we achieve speed gains of more than 100× compared to RANSAC strategies. An extensive experimental evaluation will show that our solution yields accurate results in situations with up to 50% of outliers, and can process more than 1000 correspondences in less than 5ms., This work has been partially funded by Spanish government under projects DPI2011-27510, IPT-2012-0630-020000, IPT-2011-1015-430000 and CICYT grant TIN2012-39203; by the EU project ARCAS FP7-ICT-2011-28761; and by the ERA-Net Chistera project ViSen PCIN-2013-047, Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

A high performance CRF model for clothes parsing

  • Simo-Serra, Edgar
  • Fidler, Sanja
  • Moreno-Noguer, Francesc
  • Urtasun, Raquel
Trabajo presentado a la 12th Asian Conference on Computer Vision celebrada en Singapur del 1 al 5 de noviembre de 2014.-- Publicado por Springer en "Computer Vision -- ACCV 2014 Volume 9005 of the series Lecture Notes in Computer Science pp 64-81.", In this paper we tackle the problem of clothing parsing: Our goal is to segment and classify different garments a person is wearing. We frame the problem as the one of inference in a pose-aware Conditional Random Field (CRF) which exploits appearance, figure/ground segmentation, shape and location priors for each garment as well as similarities between segments, and symmetries between different human body parts. We demonstrate the effectiveness of our approach on the Fashionista dataset and show that we can obtain a significant improvement over the state-of-the-art., This work has been partially funded by Spanish Ministry of Economy and Competitiveness under projects PAU+ DPI2011-27510 and ERA-Net Chistera project ViSen PCIN-2013-047., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Evaluation of random forests on large-scale classification problems using a bag-of-visual-words representation

  • Solé, Xavier
  • Ramisa, Arnau
  • Torras, Carme
Trabajo presentado a la 2nd KES International Conference on Innovation in Medicine and Healthcare (InMed-14), celebrada en San Sebastian (España) del 9 al 11 de julio de 2014., Random Forest is a very efficient classification method that has shown success in tasks like image segmentation or object detection, but has not been applied yet in large-scale image classification scenarios using a Bag-of-Visual-Words representation. In this work we evaluate the performance of Random Forest on the ImageNet dataset, and compare it to standard approaches in the state-of-the-art., This research is partially funded by the Spanish Ministry of Science and Innovation under project DPI2011-27510, by the CSIC project CINNOVA (201150E088) and by the ERA-Net CHISTERA project ViSen PCIN-2013-047. A. Ramisa worked under CSIC/FSE JAE-Doc grant., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Fast online learning and detection of natural landmarks for autonomous aerial robots

  • Villamizar, Michael
  • Sanfeliu, Alberto
  • Moreno-Noguer, Francesc
Presentado al ICRA 2014 celebrado en Hong Kong del 31 de mayo al 7 de junio., We present a method for efficiently detecting natural landmarks that can handle scenes with highly repetitive patterns and targets progressively changing its appearance. At the core of our approach lies a Random Ferns classifier, that models the posterior probabilities of different views of the target using multiple and independent Ferns, each containing features at particular positions of the target. A Shannon entropy measure is used to pick the most informative locations of these features. This minimizes the number of Ferns while maximizing its discriminative power, allowing thus, for robust detections at low computational costs. In addition, after offline initialization, the new incoming detections are used to update the posterior probabilities on the fly, and adapt to changing appearances that can occur due to the presence of shadows or occluding objects. All these virtues, make the proposed detector appropriate for UAV navigation. Besides the synthetic experiments that will demonstrate the theoretical benefits of our formulation, we will show applications for detecting landing areas in regions with highly repetitive patterns, and specific objects under the presence of cast shadows or sudden camera motions., This work was partially supported by the projects PAU+DPI2011-27510, RobTaskCoop DPI2010-17112, ERA-Net Chistera project ViSen PCIN-2013-047, and by the EU project ARCAS FP7-ICT-2011-28761., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Learning RGB-D descriptors of garment parts for informed robot grasping

  • Ramisa, Arnau
  • Alenyà, Guillem
  • Alenyà, Guillem
  • Moreno-Noguer, Francesc
  • Torras, Carme
Robotic handling of textile objects in household environments is an emerging application that has recently received considerable attention thanks to the development of domestic robots. Most current approaches follow a multiple re-grasp strategy for this purpose, in which clothes are sequentially grasped from different points until one of them yields a desired configuration. In this work we propose a vision-based method, built on the Bag of Visual Words approach, that combines appearance and 3D information to detect parts suitable for grasping in clothes, even when they are highly wrinkled. We also contribute a new, annotated, garment part dataset that can be used for benchmarking classification, part detection, and segmentation algorithms. The dataset is used to evaluate our approach and several state-of-the-art 3D descriptors for the task of garment part detection. Results indicate that appearance is a reliable source of information, but that augmenting it with 3D information can help the method perform better with new clothing items., This research is partially funded by the Spanish Ministry of Science and Innovation under Project PAU+ DPI2011-2751, the EU Project IntellAct FP7-ICT2009-6-269959 and the ERA-Net Chistera Project ViSen PCIN-2013-047. A. Ramisa worked under the JAE-Doc grant from CSIC and FSE., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Engineering Multiuser Museum Interactives for Shared Cultural Experiences

  • Confalonieri, Roberto
  • Yee-King, Matthew
  • Hazelden, Katina
  • D'Inverno, Mark
  • De Jonge, Dave
  • Sierra, Carles
  • Osman, Nardine
  • Agmoud, Leila
  • Prade, Henri
Multiuser museum interactives are computer systems installed in museums or galleries which allow several visitors to interact together with digital representations of artefacts and information from the museum׳s collection. In this paper, we describe WeCurate, a socio-technical system that supports co-browsing across multiple devices and enables groups of users to collaboratively curate a collection of images, through negotiation, collective decision making and voting. The engineering of such a system is challenging since it requires to address several problems such as: distributed workflow control, collective decision making and multiuser synchronous interactions. The system uses a peer-to-peer Electronic Institution (EI) to manage and execute a distributed curation workflow and models community interactions into scenes, where users engage in different social activities. Social interactions are enacted by intelligent agents that interface the users participating in the curation workflow with the EI infrastructure. The multiagent system supports collective decision making, representing the actions of the users within the EI, where the agents advocate and support the desires of their users e.g. aggregating opinions for deciding which images are interesting enough to be discussed, and proposing interactions and resolutions between disagreeing group members. Throughout the paper, we describe the enabling technologies of WeCurate, the peer-to-peer EI infrastructure, the agent collective decision making capabilities and the multi-modal interface. We present a system evaluation based on data collected from cultural exhibitions in which WeCurate was used as supporting multiuser interactive., This work is supported by the Collective Mind project (Spanish Ministry of Economy and Competitiveness, under Grant number TEC2013-49430-EXP), the PRAISE project (funded by the European Commission, under Grant number 388770), and the ACE project (funded by the European framework ERA-Net CHIST-ERA, under contract CHRI-001-03), Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

A quantitative analysis of dressing dynamics for robotic dressing assistance

  • Chance, Greg
  • Jevtić, Aleksandar
  • Caleb-Solly, Praminda
  • Dogramadzi, Sanja
Assistive robots have a great potential to address issues related to an aging population and an increased demand for caregiving. Successful deployment of robots working in close proximity with people requires consideration of both safety and human–robot interaction (HRI). One of the established activities of daily living where robots could play an assistive role is dressing. Using the correct force profile for robot control will be essential in this application of HRI requiring careful exploration of factors related to the user’s pose and the type of garments involved. In this paper, a Baxter robot was used to dress a jacket onto a mannequin and human participants considering several combinations of user pose and clothing type (base layers), while recording dynamic data from the robot, a load cell, and an IMU. We also report on suitability of these sensors for identifying dressing errors, e.g., fabric snagging. Data were analyzed by comparing the overlap of confidence intervals to determine sensitivity to dressing. We expand the analysis to include classification techniques such as decision tree and support vector machines using k-fold cross-validation. The 6-axis load cell successfully discriminated between clothing types with predictive model accuracies between 72 and 97%. Used independently, the IMU and Baxter sensors were insufficient to discriminate garment types with the IMU showing 40–72% accuracy, but when used in combination this pair of sensors achieved an accuracy similar to the more expensive load cell (98%). When observing dressing errors (snagging), Baxter’s sensors and the IMU data demonstrated poor sensitivity but applying machine learning methods resulted in model with high predicative accuracy and low false negative rates (≤5%). The results show that the load cell could be used independently for this application with good accuracy but a combination of the lower cost sensors could also be used without a significant loss in precision, which will be a key element in the robot control architecture for safe HRI., This research was supported by EPSRC and EU CHIST-ERA, UK EPSRC funding Project Reference: EP/N021703/1. This work was partially funded by the programme “Acciones de Programación Conjunta Internacional 2015” by the Spanish Ministry of Economy, Industry and Competitiveness (project reference num. PCIN-2015-147)., Peer reviewed, Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

What's "up"? - Resolving interaction ambiguity through non-visual cues for a robotic dressing assistant

  • Chance, Greg
  • Caleb-Solly, Praminda
  • Jevtić, Aleksandar
  • Dogramadzi, Sanja
Trabajo presentado al 26th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN), celebrado en Lisboa (Portugal) del 28 de agosto al 1 de septiembre de 2017., Robots that can assist in activities of daily living (ADL) such as dressing assistance, need to be capable of intuitive and safe interaction. Vision systems are often used to provide information on the position and movement of the robot and user. However, in a dressing context, technical complexity, occlusion and concerns over user privacy pushes research to investigate other approaches for human-robot interaction (HRI). We analysed verbal, proprioceptive and force feedback from 18 participants during a human-human dressing experiment where users received dressing assistance from a researcher mimicking robot behaviour. This paper investigates the occurrence of deictic speech in an assisted-dressing task and how any ambiguity could be resolved to ensure safe and reliable HRI. We focus on one of the most frequently occurring deictic words >up>, which was captured over 300 times during the experiments and is used as an example of an ambiguous command. We attempt to resolve the ambiguity of these commands through predictive models. These models were used to predict end effector choice and the direction in which the garment should move. The model for predicting end effector choice resulted in 70.4% accuracy based on the user's head orientation. For predicting garment direction, the model used the angle of the user's arm and resulted in 87.8% accuracy. We also found that additional categories such as the starting position of the user's arms and end-effector height may improve the accuracy of a predictive model. We present suggestions on how these inputs may be attained through non-visual means, for example through haptic perception of end-effector position, proximity sensors and acoustic source localisation., This research was supported by EPSRC and EU CHIST-ERA., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Robot motion adaptation through user intervention and reinforcement learning

  • Jevtić, Aleksandar
  • Colomé, Adrià
  • Alenyà, Guillem
  • Torras, Carme
Assistant robots are designed to perform specific tasks for the user, but their performance is rarely optimal, hence they are required to adapt to user preferences or new task requirements. In the previous work, the potential of an interactive learning framework based on user intervention and reinforcement learning (RL) was assessed. The framework allowed the user to correct an unfitted segment of the robot trajectory by using hand movements to guide the robot along a corrective path. So far, only the usability of the framework was evaluated through experiments with users. In the current work, the framework is described in detail and its ability to learn from a set of sample trajectories using an RL algorithm is analyzed. To evaluate the learning performance, three versions of the framework are proposed that differ in the method used to obtain the sample trajectories, which are: human-guided learning, autonomous learning, and combined human-guided with autonomous learning. The results show that the combination of the human-guided and autonomous learning achieved the best performance, and although it needed a higher number of sample trajectories than the human-guided learning, it required less user involvement. Autonomous learning alone obtained the lowest reward value and needed the highest number of sample trajectories., This work was performed as a part of the I-DRESS project (EU CHIST-ERA 2014) and funded by the programme “Acciones de Programación Conjunta Internacional 2015” of the Spanish Ministry of Economy, Industry and Competitiveness (ref. num. PCIN-2015-147), and partially supported by the Spanish national project RobInstruct (ref. num. TIN2014-58178-R)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

What's “up”? Resolving interaction ambiguity through non-visual cues for a robotic dressing assistant

  • CHANCE, GREG
  • Caleb-Solly, Praminda
  • Jevtic, Aleksandar
  • Dogramadzi, Sanja
@2017 Personal use of these materials is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating news collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, Robots that can assist in activities of daily living (ADL) such as dressing assistance, need to be capable of intuitive and safe interaction. Vision systems are often used to provide information on the position and movement of the robot and user. However, in a dressing context, technical complexity, occlusion and concerns over user privacy pushes research to investigate other approaches for human-robot interaction (HRI). We analysed verbal, proprioceptive and force feedback from 18 participants during a human-human dressing experiment where users received dressing assistance from a researcher mimicking robot behaviour. This paper investigates the occurrence of deictic speech in an assisted-dressing task and how any ambiguity could be resolved to ensure safe and reliable HRI. We focus on one of the most frequently occurring deictic words “up”, which was captured over 300 times during the experiments and is used as an example of an ambiguous command. We attempt to resolve the ambiguity of these commands through predictive models. These models were used to predict end effector choice and the direction in which the garment should move. The model for predicting end effector choice resulted in 70.4% accuracy based on the user's head orientation. For predicting garment direction, the model used the angle of the user's arm and resulted in 87.8% accuracy. We also found that additional categories such as the starting position of the user's arms and end-effector height may improve the accuracy of a predictive model. We present suggestions on how these inputs may be attained through non-visual means, for example through haptic perception of end-effector position, proximity sensors and acoustic source localisation., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Very fast solution to the PnP problem with algebraic outlier rejection

  • Ferraz Colomina, Luis
  • Binefa, Xavier
  • Moreno-Noguer, Francesc
We propose a real-time, robust to outliers and accurate solution to the Perspective-n-Point (PnP) problem. The main advantages of our solution are twofold: first, it integrates the outlier rejection within the pose estimation pipeline with a negligible computational overhead; and second, its scalability to arbitrarily large number of correspondences. Given a set of 3D-to-2D matches, we formulate pose estimation problem as a low-rank homogeneous system where the solution lies on its 1D null space. Outlier correspondences are those rows of the linear system which perturb the null space and are progressively detected by projecting them on an iteratively estimated solution of the null space. Since our outlier removal process is based on an algebraic criterion which does not require computing the full-pose and reprojecting back all 3D points on the image plane at each step, we achieve speed gains of more than 100× compared to RANSAC strategies. An extensive experimental evaluation will show that our solution yields accu- rate results in situations with up to 50% of outliers, and can process more than 1000 correspondences in less than 5ms., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Learning RGB-D descriptors of garment parts for informed robot grasping

  • Ramisa Ayats, Arnau
  • Alenyà Ribas, Guillem|||0000-0002-6018-154X
  • Moreno-Noguer, Francesc
  • Torras, Carme|||0000-0002-2933-398X
Robotic handling of textile objects in household environments is an emerging application that has recently received considerable attention thanks to the development of domestic robots. Most current approaches follow a multiple re-grasp strategy for this purpose, in which clothes are sequentially grasped from different points until one of them yields a desired configuration. In this work we propose a vision-based method, built on the Bag of Visual Words approach, that combines appearance and 3D information to detect parts suitable for grasping in clothes, even when they are highly wrinkled. We also contribute a new, annotated, garment part dataset that can be used for benchmarking classification, part detection, and segmentation algorithms. The dataset is used to evaluate our approach and several state-of-the-art 3D descriptors for the task of garment part detection. Results indicate that appearance is a reliable source of information, but that augmenting it with 3D information can help the method perform better with new clothing items., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

A quantitative analysis of dressing dynamics for robotic dressing assistance

  • CHANCE, GREG
  • Jevtic, Aleksandar
  • Caleb-Solly, Praminda
  • Dogramadzi, Sanja
Assistive robots have a great potential to address issues related to an aging population and an increased demand for caregiving. Successful deployment of robots working in close proximity with people requires consideration of both safety and human–robot interaction (HRI). One of the established activities of daily living where robots could play an assistive role is dressing. Using the correct force profile for robot control will be essential in this application of HRI requiring careful exploration of factors related to the user’s pose and the type of garments involved. In this paper, a Baxter robot was used to dress a jacket onto a mannequin and human participants considering several combinations of user pose and clothing type (base layers), while recording dynamic data from the robot, a load cell, and an IMU. We also report on suitability of these sensors for identifying dressing errors, e.g., fabric snagging. Data were analyzed by comparing the overlap of confidence intervals to determine sensitivity to dressing. We expand the analysis to include classification techniques such as decision tree and support vector machines using k-fold cross-validation. The 6-axis load cell successfully discriminated between clothing types with predictive model accuracies between 72 and 97%. Used independently, the IMU and Baxter sensors were insufficient to discriminate garment types with the IMU showing 40–72% accuracy, but when used in combination this pair of sensors achieved an accuracy similar to the more expensive load cell (98%). When observing dressing errors (snagging), Baxter’s sensors and the IMU data demonstrated poor sensitivity but applying machine learning methods resulted in model with high predicative accuracy and low false negative rates (=5%). The results show that the load cell could be used independently for this application with good accuracy but a combination of the lower cost sensors could also be used without a significant loss in precision, which will be a key element in the robot control architecture for safe HRI., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

A high performance CRF model for clothes parsing

  • Simó Serra, Edgar
  • Fidler, Sanja
  • Moreno-Noguer, Francesc
  • Urtasun, Raquel
In this paper we tackle the problem of clothing parsing: Our goal is to segment and classify different garments a person is wearing. We frame the problem as the one of inference in a pose-aware Conditional Random Field (CRF) which exploits appearance, figure/ground segmentation, shape and location priors for each garment as well as similarities between segments, and symmetries between different human body parts. We demonstrate the effectiveness of our approach on the Fashionista dataset and show that we can obtain a significant improvement over the state-of-the-art., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Fast online learning and detection of natural landmarks for autonomous aerial robots

  • Villamizar Vergel, Michael Alejandro
  • Sanfeliu Cortés, Alberto|||0000-0003-3868-9678
  • Moreno-Noguer, Francesc
We present a method for efficiently detecting natural landmarks that can handle scenes with highly repetitive patterns and targets progressively changing its appearance. At the core of our approach lies a Random Ferns classifier, that models the posterior probabilities of different views of the target using multiple and independent Ferns, each containing features at particular positions of the target. A Shannon entropy measure is used to pick the most informative locations of these features. This minimizes the number of Ferns while maximizing its discriminative power, allowing thus, for robust detections at low computational costs. In addition, after offline initialization, the new incoming detections are used to update the posterior probabilities on the fly, and adapt to changing appearances that can occur due to the presence of shadows or occluding objects. All these virtues, make the proposed detector appropriate for UAV navigation. Besides the synthetic experiments that will demonstrate the theoretical benefits of our formulation, we will show applications for detecting landing areas in regions with highly repetitive patterns, and specific objects under the presence of cast shadows or sudden camera motions.
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Evaluation of random forests on large-scale classification problems using a bag-of-visual-words representation

  • Soler, Xavier
  • Ramisa Ayats, Arnau
  • Torras, Carme|||0000-0002-2933-398X
Random Forest is a very efficient classification method that has shown success in tasks like image segmentation or object detection, but has not been applied yet in large-scale image classification scenarios using a Bag-of-Visual-Words representation. In this work we evaluate the performance of Random Forest on the ImageNet dataset, and compare it to standard approaches in the state-of-the-art., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Learning RGB-D descriptors of garment parts for informed robot grasping

  • Ramisa Ayats, Arnau
  • Alenyà Ribas, Guillem
  • Moreno-Noguer, Francesc
  • Torras, Carme
Robotic handling of textile objects in household environments is an emerging application that has recently received considerable attention thanks to the development of domestic robots. Most current approaches follow a multiple re-grasp strategy for this purpose, in which clothes are sequentially grasped from different points until one of them yields a desired configuration. In this work we propose a vision-based method, built on the Bag of Visual Words approach, that combines appearance and 3D information to detect parts suitable for grasping in clothes, even when they are highly wrinkled. We also contribute a new, annotated, garment part dataset that can be used for benchmarking classification, part detection, and segmentation algorithms. The dataset is used to evaluate our approach and several state-of-the-art 3D descriptors for the task of garment part detection. Results indicate that appearance is a reliable source of information, but that augmenting it with 3D information can help the method perform better with new clothing items., Peer Reviewed, Postprint (author’s final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

What's “up”? Resolving interaction ambiguity through non-visual cues for a robotic dressing assistant

  • CHANCE, GREG
  • Caleb-Solly, Praminda
  • Jevtic, Aleksandar
  • Dogramadzi, Sanja
@2017 Personal use of these materials is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating news collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, Robots that can assist in activities of daily living (ADL) such as dressing assistance, need to be capable of intuitive and safe interaction. Vision systems are often used to provide information on the position and movement of the robot and user. However, in a dressing context, technical complexity, occlusion and concerns over user privacy pushes research to investigate other approaches for human-robot interaction (HRI). We analysed verbal, proprioceptive and force feedback from 18 participants during a human-human dressing experiment where users received dressing assistance from a researcher mimicking robot behaviour. This paper investigates the occurrence of deictic speech in an assisted-dressing task and how any ambiguity could be resolved to ensure safe and reliable HRI. We focus on one of the most frequently occurring deictic words “up”, which was captured over 300 times during the experiments and is used as an example of an ambiguous command. We attempt to resolve the ambiguity of these commands through predictive models. These models were used to predict end effector choice and the direction in which the garment should move. The model for predicting end effector choice resulted in 70.4% accuracy based on the user's head orientation. For predicting garment direction, the model used the angle of the user's arm and resulted in 87.8% accuracy. We also found that additional categories such as the starting position of the user's arms and end-effector height may improve the accuracy of a predictive model. We present suggestions on how these inputs may be attained through non-visual means, for example through haptic perception of end-effector position, proximity sensors and acoustic source localisation., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Very fast solution to the PnP problem with algebraic outlier rejection

  • Ferraz Colomina, Luis
  • Binefa, Xavier
  • Moreno-Noguer, Francesc
We propose a real-time, robust to outliers and accurate solution to the Perspective-n-Point (PnP) problem. The main advantages of our solution are twofold: first, it integrates the outlier rejection within the pose estimation pipeline with a negligible computational overhead; and second, its scalability to arbitrarily large number of correspondences. Given a set of 3D-to-2D matches, we formulate pose estimation problem as a low-rank homogeneous system where the solution lies on its 1D null space. Outlier correspondences are those rows of the linear system which perturb the null space and are progressively detected by projecting them on an iteratively estimated solution of the null space. Since our outlier removal process is based on an algebraic criterion which does not require computing the full-pose and reprojecting back all 3D points on the image plane at each step, we achieve speed gains of more than 100× compared to RANSAC strategies. An extensive experimental evaluation will show that our solution yields accu- rate results in situations with up to 50% of outliers, and can process more than 1000 correspondences in less than 5ms., Peer Reviewed, Postprint (author’s final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

A high performance CRF model for clothes parsing

  • Simó Serra, Edgar
  • Fidler, Sanja
  • Moreno-Noguer, Francesc
  • Urtasun, Raquel
In this paper we tackle the problem of clothing parsing: Our goal is to segment and classify different garments a person is wearing. We frame the problem as the one of inference in a pose-aware Conditional Random Field (CRF) which exploits appearance, figure/ground segmentation, shape and location priors for each garment as well as similarities between segments, and symmetries between different human body parts. We demonstrate the effectiveness of our approach on the Fashionista dataset and show that we can obtain a significant improvement over the state-of-the-art., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

A quantitative analysis of dressing dynamics for robotic dressing assistance

  • CHANCE, GREG
  • Jevtic, Aleksandar
  • Caleb-Solly, Praminda
  • Dogramadzi, Sanja
Assistive robots have a great potential to address issues related to an aging population and an increased demand for caregiving. Successful deployment of robots working in close proximity with people requires consideration of both safety and human–robot interaction (HRI). One of the established activities of daily living where robots could play an assistive role is dressing. Using the correct force profile for robot control will be essential in this application of HRI requiring careful exploration of factors related to the user’s pose and the type of garments involved. In this paper, a Baxter robot was used to dress a jacket onto a mannequin and human participants considering several combinations of user pose and clothing type (base layers), while recording dynamic data from the robot, a load cell, and an IMU. We also report on suitability of these sensors for identifying dressing errors, e.g., fabric snagging. Data were analyzed by comparing the overlap of confidence intervals to determine sensitivity to dressing. We expand the analysis to include classification techniques such as decision tree and support vector machines using k-fold cross-validation. The 6-axis load cell successfully discriminated between clothing types with predictive model accuracies between 72 and 97%. Used independently, the IMU and Baxter sensors were insufficient to discriminate garment types with the IMU showing 40–72% accuracy, but when used in combination this pair of sensors achieved an accuracy similar to the more expensive load cell (98%). When observing dressing errors (snagging), Baxter’s sensors and the IMU data demonstrated poor sensitivity but applying machine learning methods resulted in model with high predicative accuracy and low false negative rates (=5%). The results show that the load cell could be used independently for this application with good accuracy but a combination of the lower cost sensors could also be used without a significant loss in precision, which will be a key element in the robot control architecture for safe HRI., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Evaluation of random forests on large-scale classification problems using a bag-of-visual-words representation

  • Soler, Xavier
  • Ramisa Ayats, Arnau
  • Torras, Carme
Random Forest is a very efficient classification method that has shown success in tasks like image segmentation or object detection, but has not been applied yet in large-scale image classification scenarios using a Bag-of-Visual-Words representation. In this work we evaluate the performance of Random Forest on the ImageNet dataset, and compare it to standard approaches in the state-of-the-art., Peer Reviewed, Postprint (author’s final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Fast online learning and detection of natural landmarks for autonomous aerial robots

  • Villamizar Vergel, Michael Alejandro
  • Sanfeliu Cortés, Alberto
  • Moreno-Noguer, Francesc
We present a method for efficiently detecting natural landmarks that can handle scenes with highly repetitive patterns and targets progressively changing its appearance. At the core of our approach lies a Random Ferns classifier, that models the posterior probabilities of different views of the target using multiple and independent Ferns, each containing features at particular positions of the target. A Shannon entropy measure is used to pick the most informative locations of these features. This minimizes the number of Ferns while maximizing its discriminative power, allowing thus, for robust detections at low computational costs. In addition, after offline initialization, the new incoming detections are used to update the posterior probabilities on the fly, and adapt to changing appearances that can occur due to the presence of shadows or occluding objects. All these virtues, make the proposed detector appropriate for UAV navigation. Besides the synthetic experiments that will demonstrate the theoretical benefits of our formulation, we will show applications for detecting landing areas in regions with highly repetitive patterns, and specific objects under the presence of cast shadows or sudden camera motions., Preprint