A Mobile Application for the Image Based Ecological Information System

Millions of people visit zoos all over the world every year. An important aspect of their experience in zoological parks is the quality of information about the animals they observe, for both educational and entertainment purposes, that they can receive during their visits. The aim of the this work is to provide a tool that can enhance the quality of experience of zoo visitors, allowing them, during their visits, to get context-specific information about the variety of animals they encounter in a simple, fast, and interactive way. This thesis presents the mobile application I designed and developed to achieve the above illustrated goal, exploiting the location services of mobile devices and the image analysis and data storage tools provided by the Image Based Ecological Information System (IBEIS). The primary functionality of the application is to instantly detect individual animals, and the species they belong to, from pictures taken by users while they are visiting a zoo. I propose a method to process the output of HotSpotter, the image recognition algorithm for patterned species integrated in IBEIS, in order to obtain high speed and precision in the task of identifying, from pictures taken by zoo visitors, individual animals and their species in an unsupervised way.