Masoud Naghedolfeizi

Artificial neural network models for predicting degradation trends in system components and sensors

A prediction model based-on artificial neural network technology was developed for trend forecasting of a given degradation process in a system component. The model utilizes the engineering analysis of the degradation process under study with the analysis of process field data and information to predict future trend in the degradation. The neural network prediction models were applied to simulated degradation data of a typical system component. The prediction results showed that the neural network models were capable of recognizing the correct future degradation trends in data even with a limited amount of input data. In addition, the models were able to capture the dynamics and nonlinearities associated with the degradation process data.

Performance analysis of a high-end CPU under a heavy computational load and varying RAM amount using thermal imaging techniques

The purpose of this research was to investigate the effects of different RAM amounts on the CPU surface temperature distribution while operating under a heavy computational load. Two experiments corresponding to different RAM amounts were carried out under relatively poor cooling conditions. A personal computer with a high-end Intel Pentium IV CPU was used to conduct the experiments. The computational load was the multiplication of two dimensional matrices (6000 by 6000) containing double precision numbers. In order to monitor temperature distributions, a high resolution thermal imaging camera was utilized. A data acquisition system was interfaced with the camera to acquire thermal images of CPU at pre-defined computational stages. The analysis of the thermal images showed that the CPU surface temperature increases as more RAM becomes available to perform the computation. The surface temperature distribution was used to identify the location of possible hotspots on the CPU

Visualizing conductive and convective heat transfer using thermographic techniques

This work explores the educational applications of thermography in teaching conductive and convective heat transfer concepts to undergraduate students. Thermography helps students effectively visualize heat transfer phenomena particularly in two and three dimensions. Experiments were designed to demonstrate one-dimensional heat transfer in metal rods, two dimensional heat transfer in a metal plate subject to given boundary conditions and convective heat transfer in fluid. The data obtained through these experiments, including thermal imaging video clips, was made available to students enrolled in instrumentation and physics courses. Using these thermographic data, students could physically observe heat transfer phenomena that otherwise would have been difficult to visualize.

Effect of RAM amount on the thermal behavior of CPU operating under a heavy computational load

The purpose of this research was to investigate the effects of different RAM amount and fan failure on bulk CPU temperature rises while operating under a heavy computational load. Two sets of experiments, each with varying amounts of RAM were carried out under CPU cooling fan-on and fan-off conditions. A personal computer with an Intel Pentium III CPU was used to conduct the experiments. The computational load was the multiplication of two dimensional matrices (3100 by 3100) containing double precision numbers. To monitor temperature rises, sensitive thermocouples were installed on the CPU heat-sink, the RAM module, and the hard disk. This paper demonstrates that CPU temperature increases as more RAM becomes available to perform the computation. The authors hypothesize that the increase in CPU temperature is correlated to full CPU utilization to carry out a heavy computational load.

Operating, monitoring and controlling plant components over cyberspace

In recent years, nearly every industry has increasingly implemented computer based measurement, instrumentation and automation technologies to control, operate, and/or monitor various plant components of industrial equipment. This has also resulted in a paradigm shift from analog to digital technologies that are suitable for communications over the Internet, Web or networked computer systems. This paper presents a methodology for remote operation and monitoring of plant components through the Internet/Web. The Internet interfacing technologies have been examined through an experimental setup used at Fort Valley State University to perform remote experiments via the Internet. The setup is a motor-generator station that can be fully operated, monitored and controlled by computer systems using Virtual Instrument programs written in LabVIEW. It also features on-line capabilities that allow users to fully operate and monitor it remotely through the Internet. The paper also addresses typical technological concerns and challenges regarding safety and security measures as well as real-time operation.

Detail preservation of morphological operations through image scaling

Morphological techniques probe an image with a structuring element. By varying the size and the shape of structuring elements, geometrical information of different parts of an image and their interrelation can be extracted for the applications of demodulating boundary, identifying components or removing noise. While large size elements benefits eliminating noise, they may be disadvantageous for preserving details in an image. Taking this into consideration, in this paper, we propose an image scaling method that will preserve detailed information when applying morphological operations to remove noise. First, a binary image is obtained, from which a Preservation Ratio Scalar (PRS) is calculated. The PRS is used for upscaling the image before morphological operations, which aims at preserving structural fine details otherwise eliminated in the original image. Finally, the morphological operator processed image is downscaled using the PRS. Experiments of target detection demonstrated the effectiveness of the proposed method in preserving the structural details such as edges while eliminating noises.

Seeded Laplacian in sparse subspace for hyperspectral image classification

Sparse Representation (SR) has received an increasing amount of interest in recent years. It aims to find the sparsest representation of each data capturing high-level semantics among the linear combinations of the base sets in a given dictionary. In order to further improve the classification performance, the joint SR that incorporates interpixel correlation information of neighborhoods has been proposed for image pixel classification. However, joint SR method yields high computational cost. To improve the performance and computation efficiency of SR and joint SR, we propose a seeded Laplacian based on sparse representation (SeedLSR) framework for hyperspectral image classification, where a hypergraph Laplacian explicitly takes into account the local manifold structure of the hyperspectral pixel in a spatial-type weighted graph. Given the training data in a dictionary, SeedLSR algorithm firstly finds the sparse representation of hyperspectral pixels, which is used to define the spectral-type affinity matrix of an undirected graph. Then, using the training data as user-defined seeds, the final classification can be obtained by solving the combination of spectral and spatial hypergraph Laplacian quadratic problem. To assess the efficiency of the proposed SeedLSR method, experiments were performed on the scene data under daylight illumination. Compared with SR algorithm, the classification results vary smoothly along the geodesics of the data manifold.

An Improved Random Walker with Bayes Model for Volumetric Medical Image Segmentation

Random walk (RW) method has been widely used to segment the organ in the volumetric medical image. However, it leads to a very large-scale graph due to a number of nodes equal to a voxel number and inaccurate segmentation because of the unavailability of appropriate initial seed point setting. In addition, the classical RW algorithm was designed for a user to mark a few pixels with an arbitrary number of labels, regardless of the intensity and shape information of the organ. Hence, we propose a prior knowledge-based Bayes random walk framework to segment the volumetric medical image in a slice-by-slice manner. Our strategy is to employ the previous segmented slice to obtain the shape and intensity knowledge of the target organ for the adjacent slice. According to the prior knowledge, the object/background seed points can be dynamically updated for the adjacent slice by combining the narrow band threshold (NBT) method and the organ model with a Gaussian process. Finally, a high-quality image segmentation result can be automatically achieved using Bayes RW algorithm. Comparing our method with conventional RW and state-of-the-art interactive segmentation methods, our results show an improvement in the accuracy for liver segmentation (p < 0.001).

Superpixel sparse representation for target detection in hyperspectral imagery

Sparse Representation (SR) is an effective classification method. Given a set of data vectors, SR aims at finding the sparsest representation of each data vector among the linear combinations of the bases in a given dictionary. In order to further improve the classification performance, the joint SR that incorporates interpixel correlation information of neighborhoods has been proposed for image pixel classification. However, SR and joint SR demand significant amount of computational time and memory, especially when classifying a large number of pixels. To address this issue, we propose a superpixel sparse representation (SSR) algorithm for target detection in hyperspectral imagery. We firstly cluster hyperspectral pixels into nearly uniform hyperspectral superpixels using our proposed patch-based SLIC approach based on their spectral and spatial information. The sparse representations of these superpixels are then obtained by simultaneously decomposing superpixels over a given dictionary consisting of both target and background pixels. The class of a hyperspectral pixel is determined by a competition between its projections on target and background subdictionaries. One key advantage of the proposed superpixel representation algorithm with respect to pixelwise and joint sparse representation algorithms is that it reduces computational cost while still maintaining competitive classification performance. We demonstrate the effectiveness of the proposed SSR algorithm through experiments on target detection in the in-door and out-door scene data under daylight illumination as well as the remote sensing data. Experimental results show that SSR generally outperforms state of the art algorithms both quantitatively and qualitatively.

Selection of principal components based on Fisher discriminant ratio

Principal component analysis transforms a set of possibly correlated variables into uncorrelated variables, and is widely used as a technique of dimensionality reduction and feature extraction. In some applications of dimensionality reduction, the objective is to use a small number of principal components to represent most variation in the data. On the other hand, the main purpose of feature extraction is to facilitate subsequent pattern recognition and machine learning tasks, such as classification. Selecting principal components for classification tasks aims for more than dimensionality reduction. The capability of distinguishing different classes is another major concern. Components that have larger eigenvalues do not necessarily have better distinguishing capabilities. In this paper, we investigate a strategy of selecting principal components based on the Fisher discriminant ratio. The ratio of between class variance to within class variance is calculated for each component, based on which the principal components are selected. The number of relevant components is determined by the classification accuracy. To alleviate overfitting which is common when there are few training data available, we use a cross-validation procedure to determine the number of principal components. The main objective is to select the components that have large Fisher discriminant ratios so that adequate class separability is obtained. The number of selected components is determined by the classification accuracy of the validation data. The selection method is evaluated by face recognition experiments.