Fantastic Beasts and How to ClassiFind Them: Pri-matrix Factorization Competition

We've partnered with Driven Data and the ARCUS foundation to offer a new €20,000 challenge that takes data scientists deep into the jungles of Africa. Camera traps are useful for non-invasive observation of wildlife and have the potential to free up huge amounts of research time -- but they can't yet automatically flag and label the species they observe.

In this project, a global community has annotated videos through the Chimp&See Zooniverse project, and now its time for the data scientist community to turn those labels into algorithms! You'll find here one of the largest labeled camera trap datasets for you to practice your skills and help researchers delve into the secrets of life on Earth!

Utilizing both crowdsourced labels as well as crowdsourced algorithms, this ambitious computer vision competition is in a league of its own. The winning techniques developed here will provide a starting point for production-level automated species tagging for use in camera trap systems around the world. By decreasing the time that experts spend watching empty footage, we can improve their ability to focus on the outcomes that matter most.

On Chimp&See people from around the world have been recording what they see in each video but they also carry out more complex tasks such as identifying unique chimpanzee individuals and tagging primates and other animals to the species level. With this competition we hope to use machine learning to accomplish the first task more time efficiently thereby allowing citizen scientists to focus on the more complex tasks of the project!

For more info check out the DrivenData blog and the competition!

New Chimp Matching Video Tutorial

As you know, the Chimp and See project includes several different activities: classifying all video clips, tagging of specific species and behaviors, and matching of individual chimps (and in the last year, leopards).  One of the comments we often hear from our newer volunteers is that they have trouble telling one chimp from another.  It does take practice, but to help speed up the learning process, the moderator team at Chimp and See would like to announce a new video tutorial!

In, "How to Recognize Chimpanzee Faces," you'll learn a step-by-step technique that shows you what to look for and how to compare individuals to find a match.  It's loosely based on a concept used by sketch artists and facial recognition software, and is simple enough for anyone to follow and put to use.  We'd love to have more of you join us in chimp matching, so if you don't have the hang of it yet, or have been hesitant to try, take a look at the new tutorial:

(14 mins., English)

When you're ready, you can practice what you've learned with the chimps on our newest site, Green Snowflake!  See you there!

Aged Violet is finished!

Aged Violet has been the largest research site we’ve worked on so far – and we finally finished it a few days ago! Thanks to everyone who was involved in the classification of nearly 100,000 clips and with chimp matching! 

A review blogpost of this site needs to start with its most discussed feature – the very many duiker videos. Those small to medium-sized antelope- or deer-like herbivores (small grey duikers aka Maxwell’s duikers, as well as the red ones: Brooke’s and Bay duikers) have been just everywhere. Even very patient and enthusiastic long-term volunteers spoke about that we should re-name the project to “DuikerWatch” as there were so many videos of them. Sometimes the duikers reacted to the camera and stretched their necks very long to see (but mostly sniff) what might be there. They are usually getting along very good with the monkey species and forage together on the forest floor.

Only three of many duikers we've seen at Aged Violet

Aged Violet also presented us the best footage of leopards of all sites. We were able to watch powerful big males on prowl in bright daylight – allowing us to identify them on the individual level in the leopard matching project. We’ve seen at least two leopard moms with a cub and a first for us – a leopard female with prey after a successful hunt. We will work furthermore on these videos in the leopard mini-project.

But the highlight of this site – well, of every site – have been the chimpanzees, of course. The Aged Violet chimp community is not very big, but has some members that were at the same time easily recognizable and outstanding as individuals. There is Filou – an older male with torn ears and scars on the back, who adopted and carries around a male infant we named Marlyn, and is seen often with his “kindergarten”: Marlyn; Kamala, another small juvenile that accompanies them almost permanently; and Sabirah, a young adolescent female is often around, too.
On the other hand, Jojo-Crann, an adolescent male, spends most of his time closely following his mom Makena who cares for a young infant. Makena’s slim figure, gray back and a damaged right ear made her easily recognizable to us. In addition, she does not like carrying her infant girl ventrally and swoops her on her back, when the little one managed to get on her belly in frightening situations, like when Makena crosses a small river.
Another individual to remember is smiling Lya. She is a young female with an enormous swelling that should be very attractive to the males here. Her broad face, the coloration of it and the form of the muzzle let her look like as if broadly smiling all the time.

We could observe the Aged Violet chimps in big traveling parties, skillful cracking nuts, and just relaxing and playing. We still have a number of open matching discussions at Aged Violet and did not name all the cute chimp infants, so your help is still wanted and maybe you get the chance to name a chimpanzee!

The project is, of course not finished with this site! Recently, we opened a new site in Eastern Africa – Green Snowflake – with new chimps and great footage of red colobus monkeys.  

Jump over to Chimp&See to help us classify African wildlife and try your hand at chimp matching!

Animal Selfies update October 2017!

Originally posted on Talk by Ammie Kalan: 

--

Hi everyone!
I’d like to take a moment to update you about the animal selfies project.
First of all, I want to thank you all for assisting me with this project by finding and tagging all camera reactions. It is a tremendous help to not only annotate the camera reactions in the classification step, but also tagging them here on Talk to allow discussions about the videos and to build tag group collections. Volunteers have already tagged almost 200 elephant camera reactions from all sites, 42 camera reactions of gorillas, more than 800 camera reactions from chimps, and many more curious baboons. This is amazing!!! We now have a solid and ever growing database for looking at how wild animals respond to camera-traps. 

As new volunteers are joining the project all the time and others might have forgotten, I want to take the opportunity to remind everyone on the purpose of this project and how you can help. I am looking into the camera reactions of chimps and gorillas (foremost) and of baboons and elephants (secondarily). I am investigating behavioral differences towards the cameras on the individual level and across different communities and sites, as well as species. The baboon and elephant data will be coded and analyzed in a second step since it is necessary to identify elephants and baboons individually (or at least to the group level) as we do it here at Chimp&See with chimps and gorillas. At the moment, this is not possible – so the apes have priority – but this will hopefully be accomplished later. Therefore, tags and videos from more sites here at Chimp&See with those four species can always be incorporated in future studies.

 I am specifically asking for your help to tag all camera reactions of these species (chimps, gorillas, baboons, and elephants) with #camera_reaction or #selfie, and to add #camtouch, if the camera is touched. An important step in this process is to also make sure that the species #chimp, #gorilla, #baboon, #elephant, and the site name, e.g., #GreenSnowflake, is tagged too.

 Thank you again for all your help! If you have any questions, please ask them here!