Automatic Microfossil Detection in Somosaguas Sur paleontologic site (Pozuelo de Alarcón, Madrid, Spain) using Multilayer Perceptrons

Microvertebrate fossils are used in biochronology to determinate the age of geological layers with a high grade of accuracy, and in paleoecology to extract information about the past enviroment. Actual techniques used to extract microfossils are manual, and require of a high amount of time and human resources. This fact make interesting the study of other more complex techniques. The work presented in this paper evaluates the use of neural networks in order to help in the microfossil extraction tasks, using a high resolution image of the sediment. Color information of the pixels are used to train Multilayer Perceptrons. The analyzed material has been obtained from Somosaguas paleontological site, in Madrid (Spain). Key-Words: Multilayer Perceptron, Neyman-Pearson, Microfossil detection, Somosaguas site.