Filters (3 )
Download Results
77 research outcomes, page 1 of 8
  • publication . Conference object . Preprint . Other literature type . 2019
    Open Access
    Authors:
    Sharib Ali; Jens Rittscher;
    Publisher: IEEE
    Project: RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    Comment: Accepted at IEEE International Symposium on Biomedical Imaging (ISBI), 2019

  • publication . Part of book or chapter of book . Preprint . 2016
    Open Access
    Authors:
    Desmaison, Alban; Bunel, Rudy; Kohli, Pushmeet; Torr, Philip H. S.; Kumar, M. Pawan;
    Publisher: Springer International Publishing
    Project: EC | HELIOS (321162), RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    Dense conditional random fields (CRF) with Gaussian pairwise potentials have emerged as a popular framework for several computer vision applications such as stereo correspondence and semantic segmentation. By modeling long-range interactions, dense CRFs provide a more d...

  • publication . Preprint . 2019
    Open Access English
    Authors:
    Zhang, Li; Li, Xiangtai; Arnab, Anurag; Yang, Kuiyuan; Tong, Yunhai; Torr, Philip H. S.;
    Project: EC | HELIOS (321162), RCUK | Understanding scenes and ... (EP/N019474/1), RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    Exploiting long-range contextual information is key for pixel-wise prediction tasks such as semantic segmentation. In contrast to previous work that uses multi-scale feature fusion or dilated convolutions, we propose a novel graph-convolutional network (GCN) to address ...

  • publication . Preprint . 2020
    Open Access English
    Authors:
    Behl, Harkirat Singh; Baydin, Atılım Güneş; Gal, Ran; Torr, Philip H. S.; Vineet, Vibhav;
    Project: EC | HELIOS (321162), RCUK | Understanding scenes and ... (EP/N019474/1), RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    Simulation is increasingly being used for generating large labelled datasets in many machine learning problems. Recent methods have focused on adjusting simulator parameters with the goal of maximising accuracy on a validation task, usually relying on REINFORCE-like gra...

  • publication . Preprint . Article . Other literature type . 2019
    Open Access English
    Authors:
    Bepler, Tristan; Morin, Andrew; Rapp, Micah; Brasch, Julia; Shapiro, Lawrence; Noble, Alex J.; Berger, Bonnie;
    Project: NIH | Structure based predictio... (2R01GM081871-05), NIH | The structural basis of c... (5R01MH114817-06), NIH | National Resource for Aut... (7P41GM103310-14), RCUK | Seebibyte: Visual Search ... (EP/M013774/1), NIH | Acquisition of Detector f... (1S10OD019994-01)

    Cryo-electron microscopy (cryoEM) is an increasingly popular method for protein structure determination. However, identifying a sufficient number of particles for analysis (often >100,000) can take months of manual effort. Current computational approaches are limited by...

  • publication . Preprint . Other literature type . Part of book or chapter of book . 2018
    Open Access English
    Authors:
    Jens Rittscher;
    Project: RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    While challenging, the dense segmentation of histology images is a necessary first step to assess changes in tissue architecture and cellular morphology. Although specific convolutional neural network architectures have been applied with great success to the problem, fe...

  • publication . Preprint . 2018
    Open Access English
    Authors:
    Gupta, Ankush; Vedaldi, Andrea; Zisserman, Andrew;
    Project: RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    End-to-end trained Recurrent Neural Networks (RNNs) have been successfully applied to numerous problems that require processing sequences, such as image captioning, machine translation, and text recognition. However, RNNs often struggle to generalise to sequences longer...

  • publication . Conference object . 2016
    Open Access English
    Authors:
    Zhong, Yujie; Arandjelović, Relja; Zisserman, Andrew;
    Publisher: HAL CCSD
    Project: EC | LEAP (336845), RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    International audience; The goal of this work is to retrieve images containing both a target person and a target scene type from a large dataset of images. At run time this compound query is handled using a face classifier trained for the person, and an image classifier...

  • publication . Preprint . 2020
    Open Access English
    Authors:
    Brown, Andrew; Xie, Weidi; Kalogeiton, Vicky; Zisserman, Andrew;
    Project: RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    Optimising a ranking-based metric, such as Average Precision (AP), is notoriously challenging due to the fact that it is non-differentiable, and hence cannot be optimised directly using gradient-descent methods. To this end, we introduce an objective that optimises inst...

  • publication . Preprint . 2018
    Open Access English
    Authors:
    Bertinetto, Luca; Henriques, João F.; Torr, Philip H. S.; Vedaldi, Andrea;
    Project: EC | HELIOS (321162), RCUK | Seebibyte: Visual Search ... (EP/M013774/1)

    Adapting deep networks to new concepts from a few examples is challenging, due to the high computational requirements of standard fine-tuning procedures. Most work on few-shot learning has thus focused on simple learning techniques for adaptation, such as nearest neighb...

77 research outcomes, page 1 of 8
Last index information