The Experts below are selected from a list of 69363 Experts worldwide ranked by ideXlab platform

Vittorio Ferrari - One of the best experts on this subject based on the ideXlab platform.

  • fluid annotation a human machine collaboration interface for full image annotation
    ACM Multimedia, 2018
    Co-Authors: Mykhaylo Andriluka, Jasper Uijlings, Vittorio Ferrari
    Abstract:

    We introduce Fluid Annotation, an intuitive human-machine collaboration interface for annotating the class label and outline of every object and background region in an image. Fluid annotation is based on three principles:(I) Strong Machine-Learning Aid. We start from the output of a strong neural network model, which the annotator can edit by correcting the labels of existing regions, adding new regions to cover missing objects, and removing incorrect regions.The edit operations are also assisted by the model.(II) Full image annotation in a single pass. As opposed to performing a series of small annotation tasks in isolation [51,68], we propose a unified interface for full image annotation in a single pass.(III) Empower the annotator.We empower the annotator to choose what to annotate and in which order. This enables concentrating on what the ma-chine does not already know, i.e. putting human effort only on the errors it made. This helps using the annotation budget effectively. Through extensive experiments on the COCO+Stuff dataset [11,51], we demonstrate that Fluid Annotation leads to accurate an-notations very efficiently, taking 3x less annotation time than the popular LabelMe interface [70].

  • fluid annotation a human machine collaboration interface for full image annotation
    arXiv: Computer Vision and Pattern Recognition, 2018
    Co-Authors: Mykhaylo Andriluka, Jasper Uijlings, Vittorio Ferrari
    Abstract:

    We introduce Fluid Annotation, an intuitive human-machine collaboration interface for annotating the class label and outline of every object and background region in an image. Fluid annotation is based on three principles: (I) Strong Machine-Learning Aid. We start from the output of a strong neural network model, which the annotator can edit by correcting the labels of existing regions, adding new regions to cover missing objects, and removing incorrect regions. The edit operations are also assisted by the model. (II) Full image annotation in a single pass. As opposed to performing a series of small annotation tasks in isolation, we propose a unified interface for full image annotation in a single pass. (III) Empower the annotator. We empower the annotator to choose what to annotate and in which order. This enables concentrating on what the machine does not already know, i.e. putting human effort only on the errors it made. This helps using the annotation budget effectively. Through extensive experiments on the COCO+Stuff dataset, we demonstrate that Fluid Annotation leads to accurate annotations very efficiently, taking three times less annotation time than the popular LabelMe interface.

Mykhaylo Andriluka - One of the best experts on this subject based on the ideXlab platform.

  • fluid annotation a human machine collaboration interface for full image annotation
    ACM Multimedia, 2018
    Co-Authors: Mykhaylo Andriluka, Jasper Uijlings, Vittorio Ferrari
    Abstract:

    We introduce Fluid Annotation, an intuitive human-machine collaboration interface for annotating the class label and outline of every object and background region in an image. Fluid annotation is based on three principles:(I) Strong Machine-Learning Aid. We start from the output of a strong neural network model, which the annotator can edit by correcting the labels of existing regions, adding new regions to cover missing objects, and removing incorrect regions.The edit operations are also assisted by the model.(II) Full image annotation in a single pass. As opposed to performing a series of small annotation tasks in isolation [51,68], we propose a unified interface for full image annotation in a single pass.(III) Empower the annotator.We empower the annotator to choose what to annotate and in which order. This enables concentrating on what the ma-chine does not already know, i.e. putting human effort only on the errors it made. This helps using the annotation budget effectively. Through extensive experiments on the COCO+Stuff dataset [11,51], we demonstrate that Fluid Annotation leads to accurate an-notations very efficiently, taking 3x less annotation time than the popular LabelMe interface [70].

  • fluid annotation a human machine collaboration interface for full image annotation
    arXiv: Computer Vision and Pattern Recognition, 2018
    Co-Authors: Mykhaylo Andriluka, Jasper Uijlings, Vittorio Ferrari
    Abstract:

    We introduce Fluid Annotation, an intuitive human-machine collaboration interface for annotating the class label and outline of every object and background region in an image. Fluid annotation is based on three principles: (I) Strong Machine-Learning Aid. We start from the output of a strong neural network model, which the annotator can edit by correcting the labels of existing regions, adding new regions to cover missing objects, and removing incorrect regions. The edit operations are also assisted by the model. (II) Full image annotation in a single pass. As opposed to performing a series of small annotation tasks in isolation, we propose a unified interface for full image annotation in a single pass. (III) Empower the annotator. We empower the annotator to choose what to annotate and in which order. This enables concentrating on what the machine does not already know, i.e. putting human effort only on the errors it made. This helps using the annotation budget effectively. Through extensive experiments on the COCO+Stuff dataset, we demonstrate that Fluid Annotation leads to accurate annotations very efficiently, taking three times less annotation time than the popular LabelMe interface.

Jasper Uijlings - One of the best experts on this subject based on the ideXlab platform.

  • fluid annotation a human machine collaboration interface for full image annotation
    ACM Multimedia, 2018
    Co-Authors: Mykhaylo Andriluka, Jasper Uijlings, Vittorio Ferrari
    Abstract:

    We introduce Fluid Annotation, an intuitive human-machine collaboration interface for annotating the class label and outline of every object and background region in an image. Fluid annotation is based on three principles:(I) Strong Machine-Learning Aid. We start from the output of a strong neural network model, which the annotator can edit by correcting the labels of existing regions, adding new regions to cover missing objects, and removing incorrect regions.The edit operations are also assisted by the model.(II) Full image annotation in a single pass. As opposed to performing a series of small annotation tasks in isolation [51,68], we propose a unified interface for full image annotation in a single pass.(III) Empower the annotator.We empower the annotator to choose what to annotate and in which order. This enables concentrating on what the ma-chine does not already know, i.e. putting human effort only on the errors it made. This helps using the annotation budget effectively. Through extensive experiments on the COCO+Stuff dataset [11,51], we demonstrate that Fluid Annotation leads to accurate an-notations very efficiently, taking 3x less annotation time than the popular LabelMe interface [70].

  • fluid annotation a human machine collaboration interface for full image annotation
    arXiv: Computer Vision and Pattern Recognition, 2018
    Co-Authors: Mykhaylo Andriluka, Jasper Uijlings, Vittorio Ferrari
    Abstract:

    We introduce Fluid Annotation, an intuitive human-machine collaboration interface for annotating the class label and outline of every object and background region in an image. Fluid annotation is based on three principles: (I) Strong Machine-Learning Aid. We start from the output of a strong neural network model, which the annotator can edit by correcting the labels of existing regions, adding new regions to cover missing objects, and removing incorrect regions. The edit operations are also assisted by the model. (II) Full image annotation in a single pass. As opposed to performing a series of small annotation tasks in isolation, we propose a unified interface for full image annotation in a single pass. (III) Empower the annotator. We empower the annotator to choose what to annotate and in which order. This enables concentrating on what the machine does not already know, i.e. putting human effort only on the errors it made. This helps using the annotation budget effectively. Through extensive experiments on the COCO+Stuff dataset, we demonstrate that Fluid Annotation leads to accurate annotations very efficiently, taking three times less annotation time than the popular LabelMe interface.

Miftachul Huda - One of the best experts on this subject based on the ideXlab platform.

  • smartphones usage in the classrooms Learning Aid or interference
    Education and Information Technologies, 2017
    Co-Authors: Muhammad Anshari, Mohammad Nabil Almunawar, Masitah Shahrill, Danang Kuncoro Wicaksono, Miftachul Huda
    Abstract:

    Many educational institutions, especially higher education institutions, are considering to embrace smartphones as part of Learning Aids in classes as most students (in many cases all students) not only own them but also are also attached to them. The main question is whether embracing smartphones in classroom teaching enhances the Learning or perhaps an interference. This paper presents the finding of our study on embracing smartphone in classroom teaching. The study was carried out through a survey and interview/discussion with a focus group of students. We found that they use their smartphones to access teaching materials or supporting information, which are normally accessible through the Internet. Students use smartphones as Learning Aids due many reasons such as they provide convenience, portability, comprehensive Learning experiences, multi sources and multitasks, and environmentally friendly. They also use smartphones to interact with teachers outside classes and using smartphones to manage their group assignments. However, integrating smartphones in a classroom-teaching environment is a challenging task. Lecturers may need to incorporate smartphones in teaching and Learning to create attractive teaching and optimum interaction with students in classes while mitigating or at least minimising distractions that can be created. Some of the challenges are distraction, dependency, lacking hands on skills, and the reduce quality of face-to-face interaction. To avoid any disturbances in using smartphones within a classroom environment, proper rules of using smartphones in class should be established before teaching, and students need to abide to these rules.

Nisarat Phithakwatchara - One of the best experts on this subject based on the ideXlab platform.

  • comparison of mannequin training satisfaction with a conventional box trainer and a low fidelity fetoscopic surgical simulator for selective fetoscopic laser photocoagulation
    Fetal Diagnosis and Therapy, 2020
    Co-Authors: Tuangsit Wataganara, Sommai Viboonchart, Wangcha Chumthup, Prakong Chuenwattana, Julaporn Pooliam, Katika Nawapun, Nisarat Phithakwatchara
    Abstract:

    Background: A low-fidelity fetoscopic surgical simulator (FSS) for training of selective fetoscopic laser photocoagulation (SFLP) was developed. Objective: To evaluate and compare training satisfaction with an FSS and with a conventional box trainer (BT). Methods: The BT consisted of a cleaned human placenta attached to the inside of a plastic storage box with a watertight lock cover and an ultrasound-transparent rubber skin. The FSS consisted of the replica of a monochorionic twin placenta attached to the inside of a spherically shaped, ultrasound-transparent phantom. Tap water was used as an ultrasound conduction agent. Evaluation of the mannequin trainings was conducted on 8 junior maternal-fetal medicine (MFM) attending physicians and 22 MFM fellows. Training satisfaction was scored from 0 to 10 on 8 different domains. Results: The mean satisfaction score (±SD) with the FSS was higher than with the BT in all domains (p < 0.05). The fellows’ training satisfaction with the BT was greater than that of the attending physicians in 4 domains: tactile feedback, demonstration of chorionic vessels, feedback on performance, and overall value as Learning Aid (p < 0.05). Conclusions: As evaluated by a small group of trainees, our FSS is superior to the BT in mannequin training of SFLP. However, the BT may be more useful for trainees with limited clinical experience.