Presentation Type

Oral Presentation

Location

Portland State University

Start Date

5-4-2016 10:00 AM

End Date

5-4-2016 11:30 AM

Subjects

Computer vision, Image analysis -- Data processing, Human-machine systems, Machine intelligence

Student Level

Undergraduate

Abstract

Image cropping is a common tool that exists in almost any image editor, yet automatic cropping is still a difficult problem in Computer Vision. Since images nowadays can be easily collected through the web, machine learning is a promising approach to solve this problem. However, an image cropping dataset is not yet available and gathering such a large-scale dataset is a non-trivial task. Although a crowdsourcing website such as Mechanical Turk seems to be a solution to this task, image cropping is a sophisticated task that is vulnerable to unreliable annotation; furthermore, collecting a large-scale high-quality dataset through crowdsourcing is expensive. Alternatively, we introduce a system that uses automatic methods and human inputs to generate and evaluate image crops. Our system is a hybrid of machine and human intelligence. Given an image, the hybrid system generates image crops in three steps: identify main objects in the image; automatically generate a set of potential good crops around the identified main objects following principle photographic composition; and assess the generated crops. The second step is automatic while the first and third steps require inputs from the human. We obtain these user inputs by designing an online game. In the user’s perspective, our system is a website where users can access to play games. In our perspective, by letting people play games, we have them annotate the images for us with no cost. The games are carefully designed so that users’ feedbacks are helpful to our main goal. The system is embedded with a quality control model that assesses the user’s accuracy and the quality of the annotation.

Rights

© Copyright the author(s)

IN COPYRIGHT:
http://rightsstatements.org/vocab/InC/1.0/
This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).

DISCLAIMER:
The purpose of this statement is to help the public understand how this Item may be used. When there is a (non-standard) License or contract that governs re-use of the associated Item, this statement only summarizes the effects of some of its terms. It is not a License, and should not be used to license your Work. To license your own Work, use a License offered at https://creativecommons.org/

Persistent Identifier

http://archives.pdx.edu/ds/psu/17230

Share

COinS
 
May 4th, 10:00 AM May 4th, 11:30 AM

Collecting Image Cropping Dataset: A Hybrid System of Machine and Human Intelligence

Portland State University

Image cropping is a common tool that exists in almost any image editor, yet automatic cropping is still a difficult problem in Computer Vision. Since images nowadays can be easily collected through the web, machine learning is a promising approach to solve this problem. However, an image cropping dataset is not yet available and gathering such a large-scale dataset is a non-trivial task. Although a crowdsourcing website such as Mechanical Turk seems to be a solution to this task, image cropping is a sophisticated task that is vulnerable to unreliable annotation; furthermore, collecting a large-scale high-quality dataset through crowdsourcing is expensive. Alternatively, we introduce a system that uses automatic methods and human inputs to generate and evaluate image crops. Our system is a hybrid of machine and human intelligence. Given an image, the hybrid system generates image crops in three steps: identify main objects in the image; automatically generate a set of potential good crops around the identified main objects following principle photographic composition; and assess the generated crops. The second step is automatic while the first and third steps require inputs from the human. We obtain these user inputs by designing an online game. In the user’s perspective, our system is a website where users can access to play games. In our perspective, by letting people play games, we have them annotate the images for us with no cost. The games are carefully designed so that users’ feedbacks are helpful to our main goal. The system is embedded with a quality control model that assesses the user’s accuracy and the quality of the annotation.