The EGI-Lifewatch Competence Centre tries to make available the EGI resource to the researchers of the Lifewatch community through three different mini-projects: Ecological observatories, Workflows & VLabs and Citizen Science. This last mini project aims to use pattern recognition tools in order to identify and classify images of different species. The last goal is to combine an citizen oriented open source app (https://www.inaturalist.org/) with a neural network based patter recognition tool (http://caffe.berkeleyvision.org/). The first approach is to train caffe to detect and classify different types of orchidees (at least in genue) using EGI resources in terms of GPGPUs. This Demo will show all the goals reached up to now and compare our method with other solutions like Mathematica Wolfram Pattern recognition.
Fernando Aguilar (CSIC)