About Us

FN5A13633

 

Core Research Team:

 uic_logo  princeton_university rpi_logo  wild-me-logo
Tanya Berger-Wolf
Computational Ecologist
University of Illinois at Chicago
USA
email Tanya
Daniel I. Rubenstein
Ecologist
Princeton University
USA
email Dan
Charles V. Stewart
Computer Vision Scientist
Rensselaer Polytechnic Institute
USA
email Chuck
Jason Holmberg
Information Architect
WildMe.org
USA
email Jason

 

 

Bios:

SONY DSC

Dr. Tanya Berger-Wolf is a Computational Ecologist, designing computational methods to solve problems in ecology, from genetics to social interactions, focusing particularly on the algorithmic and data aspects of the questions. Dr. Berger-Wolf is the Director of the IBEIS subcommittee of WildMe. She is an Associate Professor in the Department of Computer Science at the University of Illinois at Chicago, where she heads the Computational Population Biology Lab. As a legitimate part of her research (and IBEIS) she gets to fly in a super-light airplane over a nature preserve in Kenya, taking a hyper-stereo video of zebra populations and learning how to identify each one of them by the unique stripe pattern.

Dr. Berger-Wolf has received her Ph.D. in Computer Science from University of Illinois at Urbana-Champaign in 2002. After spending some time as a postdoctoral fellow working in computational phylogenetics and doing research in computational epidemiology, she returned to Illinois. She has received numerous awards for her research and mentoring, including the US National Science Foundation CAREER Award in 2008 and the UIC Mentor of the Year (2009) and Graduate Mentor (2012) awards.

rubenstein Daniel Rubenstein’s research focuses on decision-making in animals. He studies how an individual’s foraging, mating and social behavior are influenced by its phenotype, by ecological circumstances, and by the actions of other individuals in the population. He develops simple mathematical models to generate predictions that can be tested using data gathered from structured field observations or experimental manipulations. In this way he searches for general principles, or ‘rules’, that underlie complex patterns of behavior.
Much of his recent research on the adaptive value of behavior has centered on understanding the social dynamics of equids—horses, zebras and asses. In order to follow movements, social interactions and reproductive success, he needs to know every individual and monitor their actions.  And this is where IBEIS becomes an essential part of his research.

 

JasonHolmberg_headshotJason Holmberg has logged thousands of hours of development time on Wildbook and Wildbook for Whale Sharks. As Wild Me’s Information Architect, he has designed and implemented new tools to support digital pattern recognition (computer vision + artificial intelligence) for whale sharks, humpback whales, and sperm whales. Using Jason’s tool, our projects have been able to categorize and manage a large amount of wildlife data and to identify individual animals from multiple photos taken by different researchers many years apart. Jason was lead author and population modeler for two widely lauded papers covering whale shark population trajectories and Ningaloo Marine Park in Western Australia, demonstrating that citizen science data can provide improved insight and population models through greater acquisition of high quality data.

JasonPurham_headshot Qualification Summary:​

Parham is starting his 4th year as a Ph.D. student under the advising of Dr. Charles Stewart in the Computer Vision research group at RPI in Troy, NY. Before joining the Computer Vision research group, Jason spent a year studying cryptography under Dr. Bulent Yener. His current research interests are in object detection and classification using randomized techniques (e.g., Random Hough Forests) and Convolutional Neural Networks (CNNs). J​ason has visited Kenya three times since July 2014, deploying versions of the RPI research project IBEIS during each trip. His first trip deployed an IBEIS server at the Ol Pejeta Conservancy and the second trip at the Lewa Research Center, both in the Northern Laikipia region of Kenya. The third trip to Kenya, in March 2015, was to facilitate and administer the Great Zebra Count (GZC), detailed below.​  Mr. Parham also completed an internship at Kitware, which is computer vision company based in Clifton Park, NY.  During his internship with Kitware’s computer vision research group, Jason spent his time researching and contributing to analytical software for aerial and satellite imaging using CNNs.

  • Education:
    • D. (in progress) Computer Science, Rensselaer Polytechnic Institute
    • S. (in progress) Computer Science, Rensselaer Polytechnic Institute
    • S. Computer Science / Mathematics, Pepperdine University
  • Accomplishments
    • Parham’s research has been contributing to the I​BEIS i​mage analysis system that is currently being used in Nairobi National Park in Nairobi, Kenya and in choice conservancies in the northern Laikipia region of Kenya. T​he IBEIS system detects animals from images taken by trained biologists to tourists and attempts to automatically identify different individuals within populations of zebra, giraffe, elephant, and other select species.​Mr. Parham’s contributions to the project include the detection algorithms supporting plains and Grevy’s zebras, giraffes, and elephants, major advances to the interface including a web­based interface and RESTful API, and vast improvements to the underlying database storage structure.
    • In March 2015, Mr. Parham developed and administered the Great Zebra Count (GZC) in Nairobi, Kenya, which attempted to count the number of individual plain zebras and Masai giraffes in the Nairobi National Park. D​uring the GZC, 9,406 images were collected by 58 volunteer photographers driving through the park in 23 cars, each of which was assigned a preplanned route in order to ensure coverage of all parts of the park on both days. Using IBEIS, the GZC produced a two­day mark­recapture count of 2,307 +/­ 366 plains zebras and 119 +/­ 48 Masai giraffes.

 

JonVOust_headshot

Jon Van Oast has been developing online collaborative software for over twenty years.  He has a strong interest in open source software/hardware, open data, citizen science, and conservation

 

 

 

 

BlairRoberts_headshotBlair Costelloe is a behavioral ecologist specializing in the study of hoofed mammals, with lots of experience conducting fieldwork in Kenya. She has supported IBEIS in several capacities. In 2014 Blair accompanied the team to Kenya where she conducted a study on the photography patterns of safari tourists with the goal of assessing the suitability of tourists as data collectors for wildlife research projects. She also helped test early versions of the IBEIS software, provided input from a field biologist’s point of view, and provided logistical support in the field. More recently she has worked to analyze data collected during the 2014 trip and during the 2015 Great Zebra and Giraffe Count in Nairobi National Park.

 

clara

 

Clara Machogu is an ecologist with a MS degree in Range Management from University of Nairobi. She collaborated with the project, collecting data from the tourists and training the Kenyan researchers on the use of the IBEIS software

 

 

Untitled Jonathan Crall is a PhD student and the main developer of the individual recognition algorithm an is a significant contributor to the project. He has developed the QT application, the Matplotlib visualizations, and he is the main architect of the Python based API. Jonathan has made minor contributions to the web-based framework. He is currently working on improvements to the core recognition algorithm and exploring the ways that deep neural networks can be used to improve recognition performance and reduce the amount of manual verification necessary

 

marco

Marco Maggioni is a computer science researcher currently working in the High Frequency Trading area. During his PhD journey at UIC, he travelled to Kenya countless times and contributed to the IBEIS project by building and deploying the computing infrastructure use of the IBEIS software

 

 

 

Collaborators, students, contributors:
Jon Crall, computer vision PhD student, RPI
Jason Parham, computer vision PhD student, RPI
Marco Maggioni, server and network support, PhD student, UIC
Clara Machogu, IBEIS Kenya representative
Jon Van Oast, database programmer, WildMe
Michael Costelloe, web design
Roe Mary Warungu, data collection, Mpala Research Centre

The IBEIS Story


The Image Based Ecological Information System (IBEIS) platform brings a wealth of new data to science and conservation. By combining the power of big data analytics to assess wildlife health and habitat with the capability for tracking individual animals, IBEIS makes it possible to connect ecological dots in ways never before possible: to understand the dynamics of the big picture in real time.

In March 2015, IBEIS was successfully tested in Kenya, enabling the first ever Great Zebra and Giraffe Count, engaging people in a massive citizen science effort providing researchers and wildlife managers with data and an easy-to-deploy tool not only to track, but also to analyze vast amounts of previously inaccessible data. This kind of granular information is critical for the development of better, more effective resource management policies.



IBEIS is an autonomous computational system that starts with a database of photographs (including video frames) contributed by field scientists, tourists, citizen scientists, camera traps and autonomous vehicles. Designed to identify a range of species, it is also able to identify individual animals by stripes, spots, wrinkles and notches. Like fingerprints, these patterns are unique to each individual. IBEIS can find matches within the database: once an animal has been identified, it can be tracked in other photographs. This is body recognition on a massive scale. [Learn more]

ZEBRAS & GIRAFFES COUNT (4)

IBEIS provides query tools for scientists researching population demographics, species distributions, individual interactions and movement patterns. By layering additional data sets covering everything from climate change and extreme weather to agricultural development, urbanization, deforestation, the exotic animal trade, and the spread of disease, a much more detailed and useful picture of what is happening—and why—can be constructed.

IBEIS can also be used to spark interest in ecology and conservation by non-scientists as well. For example, a family or class visiting a zoo could take a picture of zebra with a phone or tablet, then access a mobile app tied into the database that would instantly generate information on the individual animal and its relatives in the wild. Since IBEIS is constantly updated, they could check back in whenever they would like for the latest news.

A BIT OF HISTORY

IBEIS was developed through the cross-disciplinary Field Computational Ecology course jointly administered by the University of Illinois at Chicago and Princeton University held at the Mpala Research Centre in Kenya. In 2010, one of the student projects in the course was an answer to a challenge to identify individual zebra from a photograph. The answer was Stripespotter, software that functions much like retail bar code, converting stripes into a unique identifier that can be quickly “read” from a field photograph. Although the initial focus was on zebras, the team quickly realized the implications for any animal with a distinctive pattern.

To date, the course has generated fourteen peer reviewed publications, two masters’ theses and countless conference presentations by both students and faculty. Stripespotter led to Hotspotter which is able to pick up distinctive wrinkle patterns, making it possible to identify more monochromatic species such as elephant and rhino. Scientists can now take a panoramic photograph to see the big savanna picture, then drill down to track a herd or an individual. Metaphorically speaking, we can now see the forest and the trees.