Danaipong Chetchotsak

An Artificial Neural Network for Detection of Simulated Dental Caries

AbstractObjects A neural network was developed to diagnose artificial dental caries using images from a charged-coupled device (CCD)camera and intra-oral digital radiography. The diagnostic performance of this neural network was evaluated against a gold standard. Materials and methods The neural network design was the Learning Vector Quantization (LVQ) used to classify a tooth surface as sound or as having dental caries. The depth of the dental caries was indicated on a graphic user interface (GUI) screen developed by Matlab programming. Forty-nine images of both sound and simulated dental caries, derived from a CCD camera and by digital radiography, were used to ‘train’ an artificial neural network. After the ’training’ process, a separate test-set comprising 322 unseen images was evaluated. Tooth sections and microscopic examinations were used to confirm the actual dental caries status.The performance of neural network was evaluated using diagnostic test. Results The sensitivity (95%CI)/specificity (95%CI) of dental caries detection by the CCD camera and digital radiography were 0.77(0.68-0.85)/0.85(0.75-0.92) and 0.81(0.72-0.88)/0.93(0.84-0.97), respectively. The accuracy of caries depth-detection by the CCD camera and digital radiography was 58 and 40%, respectively. Conclusions The model neural network used in this study could be a prototype for caries detection but should be improved for classifying caries depth. Our study suggests an artificial neural network can be trained to make the correct interpretations of dental caries.

Improve discrimination power of serum markers for diagnosis of cholangiocarcinoma using data mining-based approach

OBJECTIVE Cholangiocarcinoma (CCA) is usually fatal because of the absence of tests for early detection and lack of effective therapy. Tumor markers with adequate diagnostic values are of clinical significance. This study is aimed to improve the diagnostic power of serum markers using the computational data mining technique to develop a combined diagnostic model that yielded the best diagnostic values for CCA. DESIGN AND METHODS Eight CCA-associated markers-carcinoembryonic antigen, carbohydrate antigen 19-9, alkaline phosphatase (ALP), and gamma glutamyl transferase, biliary-ALP, mucin5AC, CCA-associated carbohydrate antigen (CCA-CA) and CA-S27-were used as the inputs for the C4.5 decision tree classification model and the selected model was confirmed by ANN analyses. Eight serum markers for CCA were determined in the training set of 85 histologically proven-CCA patients and 82 control subjects. The chosen set of combined markers that gave the best diagnostic values for CCA was then validated in the testing set of 22 CCA patients and 60 controls. RESULTS A decision tree diagram built by the C4.5 algorithm suggested the serial analysis of CCA-CA and ALP for distinguishing CCA patients from non-CCA subjects with all diagnostic parameters ≥95%. The combined tests showed a precise diagnosis in the testing set. CONCLUSIONS The C4.5 model indicates the combined markers of CCA-CA and ALP that produced the more precise diagnosis for CCA.

Combining neural networks for function approximation under conditions of sparse data: the biased regression approach

In most real world applications when large amounts of data are sparse, constructing a neural network or a committee network to achieve good performance is very difficult. This paper reports on research that attempts to improve generalization capabilities of committee networks by investigating new committee fusers under sparse data conditions. A bias/variance decomposition is performed to gain insights into the outcome. The committees are formed based on a linear combination of neural networks using the concepts of ridge regression (RR), principal component regression (PCR), and the r–k class estimator. Here, a set of trained bootstrap networks serve as an input variable and the target response is the dependent variable in the regression models. In this paper, a new algorithm—the minimum error estimate (MEE) is developed to tune the parameters of the regression models so that the committees will be combined to achieve better generalization. The experimental results demonstrate that the best of the proposed methods perform at least as well as the baseline methods under all sparse data conditions. By automatically selecting the tuning parameters, the proposed algorithms can integrate unique information from the committee members as well as effectively reduce multicollinearity effects, and thus provide reasonable performances.

Efficiency Factors for Reinforced Concrete Deep Beams: Part 2 - Code Calibration

This paper presents the limit state shear design formulas for both normal and high-strength reinforced concrete deep beams using strut-and-tie model (STM). The proposed equations are based on the STMs with six state-of-the-art efficiency factors. These STMs were improved by correcting the bias and quantifying the scatter using a Bayesian parameter estimation method. The statistical parameters of material properties, dimensions, and the accuracy of design equations are considered to develop the resistance models obtained by Monte Carlo simulations. The reliability analysis is performed to determine the strength reduction factors. The calculated values of strength reduction factors are proposed for each of the considered efficiency factors.

Efficiency Factors for Reinforced Concrete Deep Beams: Part 1 - Improved Models

A strut-and-tie model (STM) has been used as a rational and simple method for analysis and design of reinforced concrete deep beams. The STM idealizes the discontinuity regions in such members as a truss-like structure consisting of compression struts and tension ties. The strength of a deep beam is usually controlled by the capacity of the diagonal concrete struts which is generally proposed in terms of the concrete strength and the efficiency factor. The efficiency factor accounts for the variation of the concrete material and the compression softening effect. One has been used in many STMs and often exhibit biases and uncertain errors in prediction of the shear strengths of deep beams. This uncertainty is due to the imperfection of formulation, missing parameters, and insufficient experimental data. Based on a database of 406 test results and six state-of-the-art formulations of the efficiency factors found in the international building codes and literature, this paper proposes improved STMs for accurate prediction of shear strength of simply supported reinforced concrete deep beams by correcting the bias and quantifying the scatter using a Bayesian parameter estimation method.

Object's Centroid Localization Using Hu-Flusser's Moments Invariant

A combination between two components generally requires screws for assemble them together with high precision and accuracy that is need in industrial application. This research proposes the machine vision technique using Hu-Flussers moments invariant to locate centroids of target screws from a tray for loading instead of the current human vision in manual operation. To validate precision and accuracy template matching is tested in parallel with Hu-Flussers moments invariant. The results show that Hu-Flussers moments invariant is better in terms of precision and have robust ability to exclude outlier too.

Hybrid Balancing Technique Using GRSOM and Bootstrap Algorithms for Classifiers with Imbalanced Data

To deal with imbalanced data, this paper proposes a hybrid data balancing technique which incorporates both over and under-sampling approaches. This technique determines how much minority data should be grown as well as how much majority data should be reduced. In this manner, noise introduced to the data due to excessive over-sampling could be avoided. On top of that, the proposed data balancing technique helps to determine the appropriate size of the balanced data and thus computation time required for construction of classifiers would be more efficient. The data balancing technique over samples the minority data through GRSOM method and then under samples the majority data using the bootstrap sampling approach. GRSOM is used in this study because it grows new samples in a non-linear fashion and preserves the original data structure. Performance of the proposed method is tested using four data sets from UCI Machine Learning Repository. Once the data sets are balanced, the committee of classifiers is constructed using these balanced data. The experimental results reveal that our proposed data balancing method provides the best performance.

Objects Centroid Correlation Using MATLAB's Neural Network Toolbox for Visually-Guided Object Manipulation

Since a pick-and-place task plays an important role in an automatic process, it normally requires machine vision to locate an object for grasping. This paper presents a practicable method used to visually guide an object grasping a group of small, 1.1 mm diameter, screws by using an inexpensive webcam with a resolution of 640 x 480. A basic feedforward neural network is utilized to make a fitting function which associates pixel coordinates of the camera to the physical coordinates of the robot while the method of linear least squares is used for comparison in parallel. The result from the feedforward neural network shows that fifty screws can be completely manipulated from a tray after their physical coordinates are loaded into the robot while the result from the method of linear least squares shows failure when picking two of the samples.