DLS algorithms assessment applied to the identification of minor differences in sample activity

G J MESCHINO; A L DAI PRA; M. A. GONZÁLEZ; A. G. SCANDURRA; M. TRIVI; H. J. RABAL; I. L. PASSONI; M. N. GUZMAN; N. L. CAP

Pucon

Congreso; IX Reunión Iberoamericana de Óptica y XII Reunión Iberoamericana de Óptica, Láseres y Aplicaciones (RIAO/OPTILAS); 2016

Red Iberoamericana de Óptica

Dynamic laser speckle (DLS) techniques are widely used in many applications in biology, biomedicine, industry, etc.Signals generated by the intensity changes in each pixel through the speckle sequence are processed for identifying underlying activity in each point. The acquired signal is processed with differentalgorithms previously developed using multiple properties of speckle patterns and digital image processing techniques.Recently we compared the performance of dynamic speckle descriptors based in time and frequency analysis[1]. To perform this evaluation a numerical simulation was proposed to explore their linearity,robustness, sensitivity related to the samples quantity, as well as also by their computing time.In this paper we analyze the performance of different algorithms used to describe DLS image activity through the segmentation of the activity regions. In this case, both numerical simulations and an actualexperiment were performed. Then, images containing different activity states were analyzed to test the algorithms. Several sets of images were constructed with different speckle activities to compare theproperties of the algorithm identification under different conditions. In particular, we have studied the ability of the descriptors for discrimination of regions with slight differences in the sample activity.Finally we explore the use of multi-descriptors to improve the results obtained with an individual descriptor.