An Exhaustive Feature Selection Approach for Human Blastocyst Differentiation

VINCIGUERRA N; DE LA ROSA E; MARISA MARTINELLI; VELASCO JG; ANDUAGA MARCHETTI IVÁN; MONALDI N; FRESNO C; FERNÁNDEZ EA

Conferencia; 37 Anual Conference of the International Society for Clinical Biostastics; 2016

The characterization and identification of embryo development (blastocysts from no-blastocysts) may impact the following procedure of implantation for pregnancy development is crucial for in-vitro fertilization procedures. Current blastocyst selection methods are based on visual morphological characterization. In order to produce an objective characterization and selection of them, several techniques are currently under research, being one of them Fourier Transform Infrared Spectroscopy (FTIR). Traditional exhaustive feature selection algorithm (EFSA) approaches try to span all the variables combinations make it them computationally intractable in a serial environment, spite they can find the most promising set of informative variables. However, the availability of modern distributed/parallel systems allows to explore such approaches.