Overview
A capsule network provides an effective way to model
part-to-whole relationships between entities and allows to
learn viewpoint invariant representations. Through this
improved representation learning, capsule networks are able to
achieve good performance in multiple domains with a drastic
decrease in the number of parameters. Recently, capsule
networks have shown state-of-the-art results for human action
localization in a video, object segmentation in medical
images, and text classification. This tutorial will provide a
basic understanding of capsule network, and we will discuss
its use in a variety of computer vision tasks such as image
classification, object segmentation, and activity detection.
Contact: Feel free to send your queries to
yogesh[at]ucf.edu
Time and Venue
Date: Sunday, June 16, 2019
Time: 1:00 pm - 5:00 pm
Location: 104C
Schedule
- 1:00 pm - 1:05 pm
Opening remarks [Mubarak Shah, UCF] - 1:05 pm - 1:45 pm
Introduction to capsules [Sara Sabour, Google] - 1:45 pm - 2:15 pm
Capsule networks: a survey [Yogesh Rawat, UCF] - 2:15 pm - 3:05 pm
Generalization to video capsules [Kevin Duarte, UCF] - 3:05 pm - 3:15 pm
Q&A - 3:15 pm - 3:35 pm
Break - 3:35 pm - 3:50 pm
Segmentation with capsule network [Rodney Lalonde, UCF]
- 3:50 pm - 4:05 pm
3D point-capsule network [Tolga Birdal, Stanford]
- 4:05 pm - 4:45 pm
Subspace Capsule Network [Marzieh Edraki, UCF] - 4:45 pm - 5:00 pm
Q&A