Learning without Labeling for Visual Applications
April 22ed, 2022 2:10 PM-4:00PM
EE Building, B1, NCKU
Please sign up here.
Abstract
Supervised training with deep Convolutional Neural Networks (CNNs) has achieved great success in various visual recognition tasks. However, it requires large amount of well-annotated data. Data labeling, especially for large-scale image dataset, is very expensive. How to learn an effective network without the need of training data labeling has become an important problem for many applications.
A promising solution is to create a learning protocol for the neural networks, so that the neural networks can learn to teach itself without manual labels. This technique is referred as the self-supervised learning, which has recently drawn an increasing attention for improving the learning performance. In this talk, we present three examples of our work on learning effective networks without manual labeling for visual applications including fast image retrieval, multi-person parts segmentation, and 3D human mesh construction from a single image. Experimental results show the effectiveness of the proposed approaches.
Bio
Ming-Ting Sun is a Professor Emeritus at the University of Washington. Previously, he was a Director at Bellcore. He has published about 300 papers and 17 book chapters on multimedia technologies. He was an Editor-in-Chief of the Journal of Visual Communication and Image Representation (JVCI) from 2012 to 2016, the Editor-in-Chief of IEEE Transactions on Multimedia (TMM) from 2000 to 2001, and the Editor-in-Chief of the IEEE Transactions on Circuits and Systems for Video Technology (TCSVT) from 1995 to 1997. He has co-edited a book (Compressed Video over Networks) and guest-edited 12 special issues for various journals. He was a Distinguished Lecturer of IEEE Circuits and Systems Society (CASS) from 2000 to 2001 and received an IEEE CASS Golden Jubilee Medal in 2000. He served as a General Co-Chair of ICME 2016 and VCIP (Visual Communication and Image Processing) 2000, and an Honorary Chair of IEEE VCIP 2015. He is a Life Fellow of IEEE.