Vision with
Biased or Scarce Data

(VBSD)

In Conjunction with European Conference on Computer Vision 2022

2pm - 6pm IDT on October 24, 2022

About VBSD 2022

About VBSD 2022

With the increasing appetite for data in data-driven methods, the issues of biased and scarce data have become a major bottleneck in developing generalizable and scalable computer vision solutions, as well as effective deployment of these solutions in real-world scenarios. To tackle these challenges, researchers from both academia and industry must collaborate and make progress in fundamental research and applied technologies. The organizing committee and keynote speakers of VBSD 2022 consist of experts from both academia and industry with rich experiences in designing and developing robust computer vision algorithms and tranferring them to real-world solutions. VBSD 2022 provides a focused venue to discuss and disseminate research related to bias and scarcity topics in computer vision.

Where

Virtual Workshop (Details TBD)

When

2pm - 6pm IDT on Monday, October 24, 2022

Keynote Speakers

Haibin Ling

Haibin Ling

Stony Brook University

Leonid Sigal

Leonid Sigal

University of British Columbia

Patrick Pérez

Patrick Pérez

Valeo.ai

VBSD 2022 Schedule

[in IDT]

Kuan-Chuan Peng

Opening Remarks Kuan-Chuan Peng

Patrick Perez

Keynote Patrick Pérez

Frugal ML for Autonomous Driving

Paper Presentation Ioanna Gkartzonika, Nikolaos Gkalelis, Vasileios Mezaris

Learning Visual Explanations for DCNN-based Image Classifiers Using an Attention Mechanism

Paper Presentation Nguyen P. Nguyen, Ramakrishna Surya, Matthew Maschmann, Prasad Calyam, Kannappan Palaniappan, Filiz Bunyak

Self-Supervised Orientation-Guided Deep Network for Segmentation of Carbon Nanotubes in SEM Imagery

Haibin Ling

Keynote Haibin Ling

Prior Knowledge Guided Unsupervised Domain Adaptation

Break

Leonid Sigal

Keynote Leonid Sigal

Data Efficiency and Bias in Detailed Scene Understanding

Paper Presentation Jiazhi Li, Wael Abd-Almageed

CAT: Controllable Attribute Translation for Fair Facial Attribute Classification

Paper Presentation Jiageng Zhu, Hanchen Xie, Wael Abd-Almageed

Weakly Supervised Invariant Representation Learning Via Disentangling Known and Unknown Nuisance Factor

Ziyan Wu

Closing Remarks Ziyan Wu



Submission

Submission Instructions

We welcome full paper submissions (up to 14 pages, excluding references or supplementary materials). Please submit at the
VBSD 2022 @ ECCV 2022 CMT website

The paper submissions must be in pdf format and use the official ECCV 2022 templates. All submissions must be anonymous and conform to ECCV standards for double-blind review. The accepted papers will be included in the ECCV 2022 proceedings. At least one author of each accepted submission must present the paper at the workshop.  

Submission deadline: July 22, 2022 (8:59PM CET, 11:59AM PST)  
Notification to authors: August 8, 2022
Camera ready deadline:  August 15, 2022 (8:59PM CET, 11:59AM PST)

We invite the submission of original and high-quality research papers in the topics related to biased or scarce data. Accepted work will be presented as either an oral, spotlight, or poster presentation. 


Topics

The topics for VBSD 2022 include, but are not limited to:

  • Algorithms and theories for explainable and interpretable computer vision models
  • Application-specific designs for explainable computer vision, e.g., healthcare, autonomous driving, etc.
  • Algorithms and theories for learning computer vision models under bias and scarcity.
  • Performance characterization of computer vision algorithms and systems under bias and scarcity
  • Algorithms for secure and privacy-aware machine learning for computer vision
  • Algorithms and theories for trustworthy computer vision models
  • The role of adjacent fields of study (e.g, computational social science) in mitigating issues of bias and trust in computer vision
  • Continuous refinement of computer vision models using active/online learning
  • Meta-learning models from various existing task-specific computer vision models
  • Brave new ideas to learn computer vision models under bias and scarcity


Accepted Papers


Accepted Papers

  • CAT: Controllable Attribute Translation for Fair Facial Attribute Classification.
    Jiazhi Li, Wael Abd-Almageed.

  • Weakly Supervised Invariant Representation Learning Via Disentangling Known and Unknown Nuisance Factor.
    Jiageng Zhu, Hanchen Xie, Wael Abd-Almageed.

  • Learning Visual Explanations for DCNN-based Image Classifiers Using an Attention Mechanism.
    Ioanna Gkartzonika, Nikolaos Gkalelis, Vasileios Mezaris.

  • Self-Supervised Orientation-Guided Deep Network for Segmentation of Carbon Nanotubes in SEM Imagery.
    Nguyen P Nguyen, Ramakrishna Surya, Matthew Maschmann, Prasad Calyam, Kannappan Palaniappan, Filiz Bunyak.

VBSD 2022 Venue

venue

Virtual Workshop

VBSD 2022 will be held virtually at 2pm - 6pm IDT on Monday, October 24, 2022.

Organizers

Kuan-Chuan Peng

Kuan-Chuan Peng

Mitsubishi Electric Research Laboratories (MERL)

Ziyan Wu

Ziyan Wu

UII America, Inc.



Program Committee

Sk Miraj Ahmed
Yunhao Ge
Lipeng Ke
Yizhou Wang