Ana Lucía Dai Pra

Evaluation of laser dynamic speckle signals applying granular computing

The laser dynamic speckle phenomenon is a grained and fluctuant interference produced when a laser light is reflected from an illuminated surface undergoing some kind of activity. This phenomenon allows developing practical applications of unlimited use in biology and technology for being a non-destructive process, enabling the detection of not easily observable activities, such as seeds viability, paints drying, bacteria activities, corrosion processes, food decomposition, fruits bruising, etc. Sequences of intensity images are obtained in order to evaluate the phenomena dynamics, and the signals generated by the intensity changes in each pixel through the sequence are processed with the finality of identifying underlying activity in each point. This paper offers a new methodology based on granular computing to characterize the signals dynamics within the time domain, reducing the time processing and proposing news evaluation parameters to characterize speckle patterns. The methodology is applicable to stationary and non-stationary cases, enabling to monitor the phenomenon in almost real time. Two dynamic processes are analyzed to assess the goodness of the proposed methodology: fast paint drying (non-stationary) and corn seed viability (stationary), being obtained results in agreement with the physical behaviour of the observed processes.

Dynamic laser speckle: decision models with computational intelligence techniques

This paper proposes the design of decision models with Computational Intelligence techniques using image sequences of dynamic laser speckle. These models aim to characterize the dynamic of the process evaluated through Temporal History Speckle Patterns (THSP) using a set of available descriptors. The models use those sets selected to improve its effectiveness, depending on the specific application. The techniques of computational intelligence field include using Artificial Neural Networks, Fuzzy Granular Computation, Evolutionary Computation elements such as Genetic Algorithms, among others. The results obtained in experiments such as the evaluation of bacterial chemotaxis, and the estimation of the drying time of coatings are encouraging and significantly improve those obtained using a single descriptor.

A study about dimensional change of industrial parts using fuzzy rules

Abstract This work presents the development of a fuzzy model aimed to predict the dimensional changes of actual industrial parts made in a new emerging material called austempered ductile iron (ADI). The dimensional change is an important factor that must be taken into account to improve the fabrication process of any actual part of ADI. This change is produced during a heat treatment cycle and in the factory practice, its prediction is difficult due to the high number of variables affecting it. In the present case, it was considered that the dimensional change is strongly influenced by 11 operative variables of the fabrication process, associated to the chemical composition and heat treatment temperatures and holding times. The data employed to build the model corresponds to measurements done by experts on different actual industrial parts. The knowledge acquisition about the origin of the data was necessary to understand the uncertainty factors and to adjust the parameters for obtaining a proper model. The model was developed with a Takagi–Sugeno-type structure, and its parameters adjusted with the ANFIS learning algorithm and introduced into an Expert System previously developed. It allows the accurate prediction of dimensional changes and helps in the efficient design of new industrial ADI parts.

Temporal and Spatial Fuzzy Granularity of Laser and Ultrasound Frames

The use of a method based on fuzzy granular computing to process ultrasound image stack and also ultra-sound single images is proposed. Fuzzy granularity has been employed with success in the determination of speckle laser activity, this determination is usually temporal and requires a stack of consecutive time frames of the sample. In this proposal a spatial method is also presented, based on the computing of intensity grains inside a sliding window in the image. To test the performance of the method, we apply it to corn kernels images obtained with dynamic laser speckle and a set of ultra-sound (US) images of an eye with melanoma. Temporal and spatial approaches would allow identifying regions of interest in the corn grain and benign and malignant tumors within the eye ultrasound images.