Over the last few decades, many large carnivore species have undergone drastic declines. This is especially true for cheetahs and leopards, which have seen range reductions of approximately 83% and 61%, respectively. To assess the efficacy of conservation efforts, we must be able to obtain accurate population estimates at the national, regional and range-wide scales. However, counting carnivores accurately is extremely difficult; they naturally occur at low densities, are often elusive, and move across large distances, making them difficult to find. To date, researchers have relied on cost- and resource-intensive approaches to record data about individual animals using methods such as manual visual identification from photographs or video, genetic markers in excrement or hair samples, or capture-recapture techniques. Advances in extracting discriminative mathematical features from an image, even using typically noisy field pictures, suggest fully-automatic or computer-assisted animal identification from individual coat markings could revolutionize approaches to estimating densities of patterned animals.
The purpose of the workshop is to gather together experts in mathematics, electrical engineering, computer science and biological communities to present the current state of the field, and to determine in which directions research and efforts should proceed to best address this challenging problem. The workshop will promote interdisciplinary approaches and will encourage the formation of new collaborative efforts to promote better conservation and management of patterned large carnivores.