TASK-CV 2017 - Workshop on Transferring and Adapting Source Knowledge in Computer Vision (TASK-CV) 2017

ICCV workshop on Transferring and Adapting Source Knowledge in Computer Vision (TASK-CV) 2017
Venice, Italy, 29 October, 2017
Workshop site: http://adas.cvc.uab.es/task-cv2017/
Challenge site: http://ai.bu.edu/visda-2017/
***
Important Dates
***
Challenge devkit/data release: see the challenge website!
Submission deadline: June 21st, 2017
Author notification: August 4th, 2017
Camera-ready due: August 14th, 2017
Workshop date: 29 October, 2017
***
Scope
***
This would be the 4th annual workshop that brings together computer vision researchers interested in domain adaptation and knowledge transfer techniques. New this year would be the proposed Domain Adaptation Challenge, see below.
A key ingredient of the recent successes in computer vision has been the availability of visual data with annotations, both for training and testing, and well-established protocols for evaluating the results. However, this traditional supervised learning framework is limited when it comes to deployment on new tasks and/or operating in new domains. In order to scale to such situations, we must find mechanisms to reuse the available annotations or the models learned from them.
***
Topics
***
Accordingly, TASK-CV aims to bring together research in transfer learning and domain adaptation for computer vision and invites the submission of research contributions on the following topics:
-TL/DA learning methods for challenging paradigms like unsupervised, and incremental or on-line learning.
-TL/DA focusing on specific visual features, models or learning algorithms.
-TL/DA jointly applied with other learning paradigms such as reinforcement learning.
-TL/DA in the era of deep neural networks (e.g., CNNs), adaptation effects of fine-tuning, regularization techniques, transfer of architectures and weights, etc.
-TL/DA focusing on specific computer vision tasks (e.g., image classification, object detection, semantic segmentation, recognition, retrieval, tracking, etc.) and applications (biomedical, robotics, multimedia, autonomous driving, etc.).
-Comparative studies of different TL/DA methods.
-Working frameworks with appropriate CV-oriented datasets and evaluation protocols to assess TL/DA methods.
-Transferring knowledge across modalities (e.g., learning from 3D data for recognizing 2D data, and heterogeneous transfer learning)
-Transferring part representations between categories.
-Transferring tasks to new domains.
-Solving domain shift due to sensor differences (e.g., low-vs-high resolution, power spectrum sensitivity) and compression schemes.
-Datasets and protocols for evaluating TL/DA methods.
This is not a closed list; thus, we welcome other interesting and relevant research for TASK-CV.
***
Domain Adaptation Challenge
***
This year we are pleased to announce an accompanying Visual Domain Adaptation challenge. Please see the challenge website ( http://ai.bu.edu/visda-2017/ ) for details, dates, and submission guidelines.
***
Best Paper Award
***
As in previous workshops, we plan to award 1-2 best papers. Sponsors will be announced.
***
Submission
***
- All submissions will be handled via the CMT website https://cmt3.research.microsoft.com/TASKCV2017/.
- The format of the papers is the same as the ICCV main conference. The contributions will consist in Extended Abstracts (EA) of 6 pages (excluding the references).
- We accept dual submissions to ICCV 2016 and TASK-CV 2017. In other words, the submission to TASK-CV 2017 should be a 6-page summary of the submission to ICCV 2017 (Authors need to indicate the difference in their ICCV camera-ready version after their paper accepted by ICCV).
- Submissions will be rejected without review if they: exceed the page limitation or violate the double-blind policy.
- Manuscript templates can be found at the main conference website: http://iccv2017.thecvf.com/submission/main_confere...
- The accepted papers will be included in the ICCV workshop collections, and also linked in the TASK-CV webpage.
***
Workshop Chairs
***
Tatiana Tommasi, University of Rome La Sapienza, Italy
Kate Saenko, Boston University, USA
Ben Usman, Boston University, USA
Xingchao Peng, Boston University, USA
Judy Hoffman, Stanford, USA
Dequan Wang, UC Berkeley, USA
Antonio M. López, Computer Vision Center & U. Autònoma de Barcelona, Spain
Wen Li – ETH Zurich, Switzerland
Francesco Orabona, Stony Brook University, USA
David Vázquez, Computer Vision Center & U. Autònoma de Barcelona, Spain