ABOUT

ABOUT THE LINC SYSTEM

The Lion Identification Network of Collaborators (LINC) is an open source system comprised of a shared database allowing lion researchers and conservationists to follow roaming lions across vast landscapes. The system employs new methodologies for sharing data across research initiatives as well as innovative search features such as facial recognition software, which uses photographs to identify individual lions.

How will LINC help?

We know that lions, especially male lions, can move large distances between populations. But we are not sure of which populations are healthy enough to send out dispersing males, how often this happens or where they reach. This activity of dispersing is one of the most important ecological processes, although it remains one of the least understood.

Why open source?
Open source software is software whose source code is available for modification or enhancement. This method of development allows a community to focus its combined energy, building on previous work and using the shared experience of many organizations. The goal of protecting the last remaining lions is a challenging problem and one that requires the combined efforts of many. Keeping this in mind, the team behind LINC has ensured that the code base is open sourced and available to the community.

With this project researchers and conservationists now have the opportunity to expand their conservation efforts and understanding of lion dispersal to efficiently encompass and help protect the full domains of the last lions in East Africa .

About the lions

 

  • In the last 50 years, it is estimated that 50% of Africa’s lions have disappeared.
  • In the last century, it is estimated that more than 80% of lion habitat has  been lost
  • Today, there are reported to be less than an estimated 30,000 lions  remaining in the wild.

Historic(2)

Current(2)

Many factors have contributed to this shocking decline, retaliatory killings after lions attack livestock is one of the primary drivers. East Africa is one of the strongholds of lion populations but with burgeoning human populations and development, lion populations are becoming increasingly isolated, leading to local extinction.

For lions to survive, it is therefore imperative that lion researchers and conservationists collaborate on a broader scale to improve connectivity between lion populations and increase their chances of survival.

Role of the LINC system:
Dispersal between areas is absolutely critical for long-term sustainability and genetic viability of lions across East Africa. Migrating individuals can rejuvenate populations where local extinction may have occurred, as well as enable a “rescue effect” in which immigrating individuals protect a dwindling local population from extinction.

LINC will allow us to better understand these migration patterns, the movement from populations and thereby enable more effective lion conservation on a broader scale. It will allow us to pinpoint and work on priority areas between various lion populations to ensure lion survival across Maasailand and East Africa.

About the technology

 

With just a photograph and an internet connection, LINC allows lion researchers to be able to share data across landscapes and borders, enabling them to more accurately monitor and track lion populations and connectivity. This greater understanding of broad scale lion populations allows more effective conservation across what remains of African lion rangelands.

LINC uses a custom web interface that allows conservationists to search, organize, access and share lion data stored in a central database.

imageSets

At the heart of LINC is a computer aided search feature that utilizes the facial features and markings of individual lions. The computer vision system currently detects lion faces and areas of interest based on a HAAR classifier. These are then processed using a Convolutional Neural Network that develops classifiers for each individual lion.

figure2HAARnormilizedConvolutionalLayer1

The facial recognition software currently uses a combination of OpenCV and the Caffe deep learning framework to develop the classifiers it is currently running on a CUDA based NVIDA GPU.

Connecting Lions — Connecting Researchers