Insight from the early stages

Aaron Mason

It has now been one month since Wildsense was launched. In this post, we will share some of the insight already generated through early use of the game.

The community

The leaderboard is led by Jayne with an amazing 2559 gems! In total, 17401 gems have been awarded.
The community is growing every day with frequent new sign ups. However, more citizen scientists are needed for us to meet our goal in analysing hundreds of thousands of photographs. Please support us by sharing the project with your friends and colleagues. If you have Twitter, please tweet the message below (click to tweet):

Tweet: Help me find and follow the last remaining wild tigers by playing this simple @getwildsense game for the iPad

Image tags

There are currently 83298 tags found from the photographs that we have in our records. Here are some of the tags that are helping us the most in classifying tigers, wild tigers and non-tiger photos:
Tiger tags: forest, chesterzoo, longleat, park, phuket, safari, bengal, bigcat, wild…
Non-tiger tags: giraffe, elephant, bird, gorilla, unihockeylangnautigers, detroittigers, baseball

Confirmed tigers

Our system has now confirmed 1015 images with tigers out of a total of 291911 images. Many more images have been identified and thrown away. I would estimate today that this experiment is approximately 2% complete – lot’s more to do!
We have identified many tiger photos and the algorithms have appeared to have found a large number of ‘domestic tigers’ in Phuket, as well as tigers from Longleat Zoo in the UK.

ome photos have a low probability calculated because the associated tags are not relevant. Your input helps us to learn more.

Some photos have a low probability calculated because the associated tags are not relevant. Your input helps us to learn more.


Here are the responses to some questions that have been asked recently:

a. How useful are the photos to researchers?
Photos that are uploaded to the web have a lot of information. This includes data about the uploader, time, location, tags and metadata, as well as the image content itself.
Individually, a photo contains useful information for wildlife tracking, Collectively, we can reveal patterns in the data, across time and locations, and benefit from the wisdom of the crowd.

b. How are tigers in the wild disambiguated from tigers in captivity?
A combination of techniques are used to classify photographs of tigers in the wild.
We use data from photos (as outlined above) so we can filter irrelevant photos. For example, we know that wild tigers only exist in some locations in the world, so we can automatically discard photos from other places.
Citizen scientists who play the Wildsense Tigers game provide information about the photo content, this includes the photo setting i.e. is the tiger in an urban environment or a zoo?

c. How are the gems used?
Gamification is a powerful mechanism that can help motivate people to spend their time on something. The Wildsense Tigers game issues gems in return for user contribution – each time the game is played a gem is awarded. Gems are totalled and users are then ranked on a leaderboard.

d. Do you have plans to work with other endangered animals?
Yes we do! Our aim is to create a platform that can be used for other wildlife. We are already in discussions about projects for other large cats and other endangered species.

What’s next?

1. We are planning an update to the Wildsense Tigers iPad game that will include many features that have been requested.
2. Soon we will put out a request for conservationist and scientists to provide us with more camera trap footage.

We plan to provide an update on progress each month with some insights into the project. If there is anything you are interested in and would like to learn more about next time please contact us or write a comment below.