[visionlist] ECCV Workshop on Transferring and Adapting Source Knowledge & VISDA Challenge

Tatiana Tommasi tommasi.t at gmail.com
Fri Jun 22 17:05:58 -05 2018

                                         2nd Call for Papers

ECCV TASK-CV Workshop on Transferring and Adapting Source Knowledge
in Computer Vision & VisDA Challenge
Munich, September 14th 2018

Workshop site: https://sites.google.com/view/task-cv2018/home
Challenge site: http://ai.bu.edu/visda-2018/

Important Dates

Paper Track
Submission: July 2​nd​, 2018
Notification: July 15​th​, 2018
Camera Readay: July 25​th​, 2018

Registration: April 21st​ , 2018
Train and Validation data release: May 16th, 2018
Test data release: August 1st, 2018
Notification win.: September 1​st​, 2018

Workshop Topics
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
protocols for evaluating the results. However, this traditional supervised
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 and generalize to
domains and tasks.

Accordingly, TASK-CV aims to bring together research in transfer learning
and domain
adaptation for computer vision and invites the submission of research
on the following topics:
■ TL/DA focusing on specific computer vision tasks (e.g., image
object detection, semantic segmentation, recognition, retrieval, tracking,
and applications (biomedical, robotics, multimedia, autonomous driving,
■ TL/DA focusing on specific visual features, models or learning algorithms
challenging paradigms like unsupervised, reinforcement, or online learning
■ TL/DA in the era of convolutional neural networks (CNNs), adaptation
of fine-tuning, regularization techniques, transfer of architectures and
weights, etc.
■ Comparative studies of different TL/DA methods and transferring part
between categories and 2D/3D modalities
■ Working frameworks with appropriate CV-oriented datasets and evaluation
protocols to assess TL/DA

This is not a closed list; thus, we welcome other related research for

VisDA Challenge
The VisDA challenge aims to test domain adaptation methods’ ability to
source knowledge and adapt it to novel target domains.

Tatiana Tommasi , IIT Milan-Italy
David Vázquez, Element AI
Kate Saenko, Boston University
Ben Usman, Boston University
Xingchao Peng, Boston University
Judy Hoffman, UC Berkeley
Neela Kaushik, Boston University
Kuniaki Saito, Boston University
Antonio M. López, UAB/CVC
Wen Li, ETH Zurich
Francesco Orabona, Boston University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20180623/1ee743a9/attachment.html>

More information about the visionlist mailing list