[visionlist] Request for information: automated image analysis methods

[Horowitz, Todd (NIH/NCI) [E]]

I received this email privately, but I thought it would be worth replying to the whole list(s).

There is a small, but growing field of researchers working on public health problems using large image collections. Sometimes these are images collected directly from participants for the purposes of the study; sometimes they are collected for the study by fixed cameras, or even drones;, sometimes they are scraped from existing collections of images on the web.

In particular, these are projects about obesity, physical activity, and the built environment. As an example, take a look at Cowburn, Gill, Anne Matthews, Aiden Doherty, Alex Hamilton, Paul Kelly, Julianne Williams, Charlie Foster, and Michael Nelson. “Exploring the Opportunities for Food and Drink Purchasing and Consumption by Teenagers during Their Journeys between Home and School: A Feasibility Study Using a Novel Method.” Public Health Nutrition 19, no. 01 (January 2016): 93–103. In this study, students in England wore cameras while going to and from school, and the researchers identified food purchase and consumption events, then checked them against GPS coordinates to see whether these events tended to take place in close proximity to the school.

I’m looking for examples of methods for automatically coding these types of data. I’m not looking for specific public health applications, because I don’t think there are many. Instead, I’d like to be able to point the public health researchers to resources from vision science that they could adapt to their specific interests.

Thanks
Todd


From: nasrin bastani [mailto:nasrin.bstn at gmail.com]
Sent: Friday, September 15, 2017 10:15 PM
To: Horowitz, Todd (NIH/NCI) [E] <todd.horowitz at nih.gov>
Subject: Re: [visionlist] Request for information: automated image analysis methods

Dear Todd

Whould you please explain more about the problem and the desired result please? so I can be sure that my solution is suitable for your research.

Best Regards
Nasrin



On 15 Sep 2017 01:41, "Horowitz, Todd (NIH/NCI) [E]" <todd.horowitz at nih.gov<mailto:todd.horowitz at nih.gov>> wrote:
Dear colleagues

I’m working on a review paper on the topic of the use of images in public health research. The research projects include collection of images from participants’ mobile phones, from wearable or stationary cameras, or images from Google Street View/Google Earth. The research questions may involve physical activity, eating/nutrition, and/or the environments that these behaviors take place in. for example, participants might be asked to take a picture of their meal before they eat it so that researchers can assess dietary patterns.

The vast majority of these studies hand-code their data. It is my belief that there should be existing automated (or perhaps crowdsourced) image analysis methods that could be adapted for these purposes. I’m hoping my vision science colleagues could point me to some useful examples.

thanks
Todd


Todd S. Horowitz, Ph.D.
Program Director, Psychologist
Basic Biobehavioral and Psychological Sciences Branch
Behavioral Research Program
Division of Cancer Control and Population Sciences
National Cancer Institute
9609 Medical Center Drive, 3E-116
Rockville, MD 20850
Phone: 240.276.6963
Mobile: 240.678.5986
[witter_450px]


_______________________________________________
visionlist mailing list
visionlist at visionscience.com<mailto:visionlist at visionscience.com>
http://visionscience.com/mailman/listinfo/visionlist_visionscience.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20170918/2ab29cad/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 25474 bytes
Desc: image001.png
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20170918/2ab29cad/attachment-0001.png>