Learning visual tasks with selective attention

Knowing where to look in an image can significantly improve performance in computer vision tasks by eliminating irrelevant information from the rest of the input image, and by breaking down complex scenes into simpler and more familiar sub-components. We show that a framework for identifying multiple task-relevant regions can be learned in current state-of-the-art deep network architectures, resulting in significant gains in several visual prediction tasks. We will demonstrate both directly and indirectly supervised models for selecting image regions and show how they can improve performance over baselines by means of focusing on the right areas.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2017

Author dc:creator

• Shih, Kevin Jonathan

Contributors dc:contributor

• Hoiem, Derek
• Lazebnik, Svetlana
• Forsyth, David
• Parikh, Devi

• Computer vision
• Visual attention
• Visual question answering (VQA)
• Keypoint localization
• Part localization
• Image recognition
• Fine-grained image recognition
• Deep learning
• Multi-task learning
• Machine learning

Statement dc:rights

• Copyright 2017 Kevin Jonathan Shih

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/98359

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/98359