Embedded Files
PROGRAM SCHEDULE
PROGRAM SCHEDULE
IW-FCV 2021 Schedule
IW-FCV 2021 Schedule
Feb. 22: main event day 1
→ local offline meeting plus Steering Committee Microsoft Teams meeting
→ Fully Online + On-Demand Video System
Feb. 23: main event day 2
→ Fully Online + On-Demand Video System
→ Award ceremony
→ Announcement for the next conference
ATTENDING
ATTENDING
Conference Instructions (Joining Online)
Conference Instructions (Joining Online)
We ask that all speakers join their live session at least 15 minutes early to ensure that there are no issues prior to the beginning of your session. Please see the format of your session below:
Regular Session Papers (Microsoft Teams session is only for Q&A):
Video presentations for regular oral and poster papers will be available as on-demand pre-recorded video. You will receive a separate email with access to the Microsoft Teams sessions.
You are expected to attend the Live Q&A Session that is assigned to your session. During the conference, presenting authors of all accepted papers are expected to attend their scheduled LIVE Q&A session (details above) where the Session Chair(s) will present a 5-min summary of the papers in the session and conduct the Q&A session. Note that the papers will NOT be presented/streamed in this LIVE Q&A session. The date and time for each LIVE Q&A session are included in the Technical Program.
Session Chairs will monitor the Chat within the Platform to make sure authors respond within a reasonable time (24 hours). The timing of the LIVE Q&A sessions is strict, so session chairs should observe the starting and end times of their session. The live Microsoft Teams session will promptly end at the scheduled time to allow for the next LIVE session to begin.
Should you have any specific requests for your live Microsoft Teams session that has yet to be discussed, please email iwfcv2021@gmail.com directly for discussion and confirmation. We look forward to your participation that makes IW-FCV2021 a success.
Detailed Presentation Schedule
Detailed Presentation Schedule
Important*: The Time mentioned in the detailed program Schedule below is according to Korea Standard Time (KST)
Notice :Click on it link to Join our Platform
Conference Day 1: FEBRUARY 22, MONDAY
Conference Day 1: FEBRUARY 22, MONDAY
Conference Day 2: FEBRUARY 23, TUESDAY
Conference Day 2: FEBRUARY 23, TUESDAY
Invited Talk (Conference Day 1)
Invited Talk (Conference Day 1)
Title: Multimodal Understanding: Vision and Language, and its Beyond
Title: Multimodal Understanding: Vision and Language, and its Beyond
Abstract: Benefited from deep learning, many computer science fields such as Computer Vision (CV) and Natural Language Processing (NLP) have experienced dramatic improvements in the existing problems such as classification, detection, and segmentation of images, and machine translation of texts. Moreover, the most significant fruit is that the typical deep learning methods reduce the barrier to entry to each other. As a result, unified understandings of multimodal data are vastly tackled. In this talk, I will introduce these challenging tasks called Vision and Language.
Abstract: Benefited from deep learning, many computer science fields such as Computer Vision (CV) and Natural Language Processing (NLP) have experienced dramatic improvements in the existing problems such as classification, detection, and segmentation of images, and machine translation of texts. Moreover, the most significant fruit is that the typical deep learning methods reduce the barrier to entry to each other. As a result, unified understandings of multimodal data are vastly tackled. In this talk, I will introduce these challenging tasks called Vision and Language.
Speaker: Dr. Yoshitaka Ushiku, PI with OMRON SINIC X and Chief Research Officer with Ridge-in
Speaker: Dr. Yoshitaka Ushiku, PI with OMRON SINIC X and Chief Research Officer with Ridge-in
Bio: Yoshitaka Ushiku is a Principal Investigator at OMRON SINIC X and Chief Research Officer at Ridge-i. He received his B.E., M.A., and Ph.D. degrees from the University of Tokyo in 2009, 2011, and 2014, respectively. In 2014, he joined NTT CS Labs, Japan, where he was involved in research on image recognition. From 2016 to 2018, he was a lecturer at the University of Tokyo, Japan. Currently, he is a Principal Investigator at OMRON SINIC X and Chief Research Officer at Ridge-i since 2018 and 2019, respectively. His research interests lie in cross-media understanding through machine learning, mainly for computer vision and natural language processing
Invited Talk (Conference Day 2)
Invited Talk (Conference Day 2)
Title: Leveraging Attention in Deep Learning for Video QA
Title: Leveraging Attention in Deep Learning for Video QA
Abstract: We are bombarded every day by an immense amount of information which our brain is incapable of processing. However, our cognitive process allows us to selectively concentrates on a discrete aspect of the information. Deep learning has incorporated a mechanism referred to as attention to partially mimic this capability. This talk will look at some attention mechanisms and focus on how it is used for video QA.
Abstract: We are bombarded every day by an immense amount of information which our brain is incapable of processing. However, our cognitive process allows us to selectively concentrates on a discrete aspect of the information. Deep learning has incorporated a mechanism referred to as attention to partially mimic this capability. This talk will look at some attention mechanisms and focus on how it is used for video QA.
Speaker: Prof. Chang D. Yoo, School of Electrical Engineering, KAIST
Speaker: Prof. Chang D. Yoo, School of Electrical Engineering, KAIST
Bio: He received the B.S. degree in Engineering and Applied Science from California Institute of Technology , the M.S. degree in Electrical Engineering from Cornell University and the Ph.D degree in Electrical Engineering from Massachusetts Institute of Technology . From January 1997 to March 1999 he was a Senior Researcher at Korea Telecom (KT). Since 1999, he has been on the faculty at the Korea Advance Institute of Science and Technology (KAIST), where he is currently a full tenured professor in the Department of Electrical Engineering and an adjunct professor in the Department of Computer Science. He is also the Associate Vice President of the Office of Special Projects and Institutional Relations.
Invited Talk (Conference Day 2)
Invited Talk (Conference Day 2)
Title: Anomalous event detection using weakly supervised learning with clustering supervision
Title: Anomalous event detection using weakly supervised learning with clustering supervision
Abstract: Real-world anomalous events detection through video-level labels is a challenging task due to rare occurrence of anomalies and noise in the labels. We propose a weakly supervised anomaly detection method which has many contributions including inter-batch correlation reduction and a normalcy suppression mechanism to minimize anomaly scores of the normal regions. In addition to that, we also introduce a clustering distance based loss to contribute towards mitigating the label noise and to produce better anomaly representations by encouraging our model to generate distinct normal and anomalous clusters. The proposed method has obtained excellent performance on publicly available datasets including UCF Crime and ShanghaiTech, compared to the existing state-of-the-art algorithms.
Abstract: Real-world anomalous events detection through video-level labels is a challenging task due to rare occurrence of anomalies and noise in the labels. We propose a weakly supervised anomaly detection method which has many contributions including inter-batch correlation reduction and a normalcy suppression mechanism to minimize anomaly scores of the normal regions. In addition to that, we also introduce a clustering distance based loss to contribute towards mitigating the label noise and to produce better anomaly representations by encouraging our model to generate distinct normal and anomalous clusters. The proposed method has obtained excellent performance on publicly available datasets including UCF Crime and ShanghaiTech, compared to the existing state-of-the-art algorithms.
References:
[1] Zaheer, Muhammad Zaigham; Lee, Jin-ha; Astrid, Marcella; Mahmood, Arif; Lee, Seung-Ik; Cleaning Label Noise With Clusters for Minimally Supervised Anomaly Detection, Computer Vision Pattern Recognition Workshop (CVPRW), 2020.
[2] Zaheer, Muhammad Zaigham; Mahmood, Arif; Shin, Hochul; Lee, Seung-Ik; A Self-Reasoning Framework for Anomaly Detection Using Video-Level Labels, IEEE Signal Processing Letters (SPL), 2020.
[3] Zaheer, Muhammad Zaigham; Mahmood, Arif; Astrid, Marcella; Lee, Seung-Ik; CLAWS: Clustering Assisted Weakly Supervised Learning with Normalcy Suppression for Anomalous Event Detection, European Conference on Computer Vision (ECCV), 2020.
Speaker: Prof. Arif Mahmood, Information Technology University (ITU), Pakistan
Speaker: Prof. Arif Mahmood, Information Technology University (ITU), Pakistan
Bio: Dr. Arif Mahmood received his MS and PhD in Computer Science from the Lahore University of Management Sciences (LUMS), winning a gold medal along the way and graduating with academic distinction. His PhD thesis was on “Computation Elimination Algorithms for Correlation Based Fast Template Matching.” He has served as a Research Assistant Professor at the School of Computer Science and Software Engineering, and later at the School of Mathematics and Statistics in the University of Western Australia from 2012-2015. Previously, he was an Assistant Professor at the Punjab University College of Information Technology from 2008-2012. He also worked as a Postdoctoral Researcher at the College of Engineering in Qatar University from 2015-2018. In 2011, Dr. Mahmood received a grant from the University of Punjab and established an image processing research facility by purchasing state-of-the-art video and still cameras, workstations and other required apparatus.
Awards
Awards
The following awards are planned:
Best paper award
Best student paper award
Best poster presentation award
Page updated
Google Sites
Report abuse