Input data for Artificial Intelligence research for the 2019 Measuring aesthetics project (NESP TWQ 5.5, Griffith Institute for Tourism Research)
- Between 01/01/2019 - 00:00 and 30/09/2020 - 00:00
The last stream within the NESP 5.5 project was related to the conduct of an online survey to get aesthetic ratings of additional 3500 images downloaded from Flickr to improve the Artificial Intelligence (AI)-based system recognising and assessing the beauty of natural scenes, which had been developed in the previous NESP 3.2.3 project. Despite some earlier investment into this research area, there is still a need to improve the tools we use to measure the aesthetic beauty of marine landscapes. This research drew on images publicly available on the Internet (in particular through the photo sharing site Flickr) to build a large dataset of GBR images for the assessment of aesthetic value. Building on earlier work in NESP TWQ Hub Project 3.2.3, we conducted a survey focused on collecting beauty scores of an additional large number of GBR images (n = 3500).
This dataset consists of one dataset report, two word files and one excel file demonstrating the aesthetic ratings collected used to improve the accuracy of the aesthetic monitoring AI system.
Methods:
The third research stream was conducted on the basis of an online survey to collect aesthetic ratings of 1585 Australians to rate the aesthetic beauty of 3500 GBR underwater pictures downloaded and selected from Flickr. Flickr is an image hosting service and one of the main sources of images for our project. As per our requirement, we downloaded all images and their metadata (including coordinates where available) based on keyword filter such as “Great Barrier Reef”. The Flickr API is available for non-commercial (but commercial use is possible by prior arrangement) use by outside developers. To ensure a much larger and diverse supply of photographs, we have developed a python-based application using Flickr API that allowed us to download Flickr images by keyword (e.g. “Great Barrier Reef” available at https://www.flickr.com). The focus of this research was on under-water images, which had to be filtered from the downloaded Flickr photos. From the collected images we identified an additional number of 3020 relevant images with coral and fish contents out of a total of approximately 55,000 downloaded images. Matt Curnock, CSIRO expert, also provide 100 images from his private images taken at the GBR and consent to use these images for our research. In total, 3120 images were selected and renamed to be rated in a survey by Australian participants (see two file “Image modification” and “Matt image rename” in the AI folder for further details).
The survey was created on Qualtrics website and launched in in April 2020 using Qualtrics survey service. After giving the consent to participating in the online survey, each respondent was randomly exposed to 50 images of the GBR and rate the aesthetic of the GBR scenery on a 10 point scale (1-Very ugly/unpleasant – Very beautiful/pleasant). In total, 1585 complete and valid questionnaires were recorded. Aesthetic rating results was exported to an Excel file and used for improving the accuracy of the computer algorithm recognising and assessing the beauty of natural scenes which had been developed in the previous NESP 3.2.3 project.
Further information can be found here:
Stantic, B. and Mandal, R. (2020) Aesthetic Assessment of the Great Barrier Reef using Deep Learning. Report to the National Environmental Science Program. Reef and Rainforest Research Centre Limited, Cairns (30pp.). Available at https://nesptropical.edu.au/wp-content/uploads/2020/11/NESP-TWQ-Project-...
Format:
The AI DATASET has one dataset report, one excel file showing aesthetic ratings of all images and two Word files showing how images downloaded from Flickr website and provided by Matt Curnock (CSIRO) were renamed and used for aesthetic ratings and AI development. The aesthetic rating results were later used to improve the accuracy of the AI aesthetic monitoring system for the GBR.
Further information can be found here:
Stantic, B. and Mandal, R. (2020) Aesthetic Assessment of the Great Barrier Reef using Deep Learning. Report to the National Environmental Science Program. Reef and Rainforest Research Centre Limited, Cairns (30pp.). Available at https://nesptropical.edu.au/wp-content/uploads/2020/11/NESP-TWQ-Project-...
References:
Murray, N., Marchesotti, M. & Perronnin, F (2012). AVA: A Large-Scale Database for Aesthetic Visual Analysis. Available (09/10/17) http://refbase.cvc.uab.es/files/MMP2012a.pdf
Data Location:
This dataset is filed in the eAtlas enduring data repository at: data\custodian\2019-2022-NESP-TWQ-5\5.5_Measuring-aesthetics
- Becken, Susanne, Professor
Griffith Institute for Tourism, Griffith University
s.becken@griffith.edu.au
- Connolly, Rod, Professor
School of Environment & Australian Rivers Institute - Coast & Estuaries, Griffith University
r.connolly@griffith.edu.au - Stantic, Bela, Professor
Head of School of Information and Communication Technology Director of "Big Data and Smart Analytics" Lab - IIIS
Griffith Sciences, Griffith University
b.stantic@griffith.edu.au - Michelle, Whitford, Associate Professor
Deputy Director of Griffith Institute for Tourism Research
Griffith University
m.whitford@griffith.edu.au - Mandal, Ranju, Dr
School of Information and Communication Technology, Griffith University
r.mandal@griffith.edu.au - Le, Dung, Dr
Griffith Institute for Tourism Research, Griffith University
d.le@griffith.edu.au, dung.ltp@vinuni.edu.vn
- Becken, Susanne, Professor
Griffith Institute for Tourism, Griffith University
s.becken@griffith.edu.au