This paper proposes a new approach for training support vector machines with a bone age determination system.The proposed approach is a combination of particle swarm optimization (PSO) and support vector machines (SVMs).The performance and accuracy of the proposed PSO-SVM algorithm are examined on a bone age data set.The results obtained by PSO-SVM show that PSO-SVM is more effective than the previous study based on conventional SVM. The evaluation of bone development is a complex and time-consuming task for the physicians since it may cause intraobserver and interobserver differences. In this study, we present a new training algorithm for support vector machines in order to determine the bone age in young children from newborn to 6 years old. By the new algorithm, we aimed to assist the radiologists so as to eliminate the disadvantages of the methods used in bone age determination. To achieve this purpose, primarily feature extraction procedure was performed to the left hand wrist X-ray images by using image processing techniques and the features related with the carpal bones and distal epiphysis of radius bone were obtained. Then these features were used for the input arguments of the classifier. In the classification process, a new training algorithm for support vector machines was proposed by using particle swarm optimization. When training support vector machines, particle swarm optimization was used for generating a new training instance which will represent the whole training set of the related class by using the training set. Finally, these new instances were used as the support vectors and classification process was carried out by using these new instances. The performance of the proposed method was compared with the naive Bayes, k-nearest neighborhood, support vector machines and C4.5 algorithms. As a result, it was determined that the proposed method was found successful than the other methods for bone age determination with a classification performance of 74.87%.

Support vector machines classification based on particle swarm optimization for bone age determination

Rheumatoid arthritis (RA) is an autoimmune disease whose common manifestation involves the slow destruction of joint tissue, a damage that is visible in a radiograph. Over time, this damage causes pain and loss of functioning, which depends, to some extent, on the spatial deformation induced by the joint damage. Building an accurate model of the current deformation and predicting potential future deformations are the important components of treatment planning. Unfortunately, this is currently a time-consuming and labor-intensive manual process. To address this problem, we propose a fully automated approach for fitting a shape model to the long bones of the hand from a single radiograph. Critically, our shape model allows sufficient flexibility to be useful for patients in various stages of RA. Our approach uses a deep convolutional neural network to extract low-level features and a conditional random field (CRF) to support shape inference. Our approach is significantly more accurate than previous work that used hand-engineered features. We provide a comprehensive evaluation for various choices of network hyperparameters, as current best practices lack significantly in this domain. We evaluate the accuracy of our pipeline on two large datasets of hand radiographs and highlight the importance of the low-level features, the relative contribution of different potential functions in the CRF, and the accuracy of the final shape estimates. Our approach is nearly as accurate as a trained radiologist and, because it only requires a few seconds per radiograph, can be applied to large datasets to enable better modeling of disease progression.

Automatic Hand Skeletal Shape Estimation From Radiographs

Rheumatoid arthritis (RA) is an autoimmune disease whose common manifestation involves the slow destruction of joint tissue, damage that is visible in a radiograph. Over time, this damage causes pain and loss of functioning which depends, to some extent, on the spatial deformation induced by the joint damage. Building an accurate model of the current deformation and predicting potential future deformations is an important component of treatment planning. Unfortunately, this is currently a time consuming and labor intensive manual process. To address this problem, we propose a fully automated approach for fitting a shape model to the long bones of the hand from a single radiograph. Critically, our shape model allows sufficient flexibility to be useful for patients in various stages of RA. Our approach uses a deep convolutional neural network to extract low-level features and a conditional random field (CRF) to support shape inference. Our approach is significantly more accurate than previous work that used hand-engineered features. We demonstrate this on two large datasets of hand radiographs and highlight the importance of the low-level features, the relative contribution of different potential functions in the CRF, and the accuracy of the final shape estimates. Our approach is nearly as accurate as a trained radiologist and, because it only requires a few seconds per radiograph, can be applied to large datasets to enable better modeling of disease progression. We will release our code and trained models upon acceptance of this paper.

Automatic Hand Skeletal Shape Estimation from Radiographs

The Tree seed algorithm (TSA) is a metaheuristic algorithm inspired by the relationship between trees and seeds. It has been proposed for very low-dimensional optimization problems and achieved promising results compared to other optimization algorithms. However, it has been determined that the performance of the TSA is lower than other algorithms for high-dimensional problems. This is due to the fact that TSA cannot scan the local optimum and search space effectively. A new TSA based on the roulette wheel strategy (R-TSA) has been proposed in this study to eliminate this disadvantage and solve high-dimensional problems. With this strategy, the trees selected at the seed production phase of TSA were diversified and the locations of the seeds were updated to prevent it from being stuck in local minima, with the aim of scanning the search space more effectively. The R-TSA was applied to high-dimensional (20, 50 and 100) benchmark functions and both convergence and box-plot graphs were obtained by using the results of these functions. Moreover, current algorithms in published literature were applied to these functions and the results obtained were compared with the R-TSA. It was observed from the analysis results that the performance of the R-TSA was higher than that of TSA.

Solving continuous optimization problems using the tree seed algorithm developed with the roulette wheel strategy

Bone age assessment (BAA) is a radiological process to identify the growth disorders in children. Although this is a frequent task for radiologists, it is cumbersome. The objective of this study is to assess the bone age of children from newborn to 18 years old in an automatic manner through computer vision methods including histogram of oriented gradients (HOG), local binary pattern (LBP), and scale invariant feature transform (SIFT). Here, 442 left-hand radiographs are applied from the University of Southern California (USC) hand atlas. In this experiment, for the first time, HOG–LBP–dense SIFT features with background subtraction are applied to assess the bone age of the subject group. For this purpose, features are extracted from the carpal and epiphyseal regions of interest (ROIs). The SVM and 5-fold cross-validation are used for classification. The accuracy of female radiographs is 73.88% and of the male is 68.63%. The mean absolute error is 0.5 years for both genders’ radiographs. The accuracy a within 1-year range is 95.32% for female and 96.51% for male radiographs. The accuracy within a 2-year range is 100% and 99.41% for female and male radiographs, respectively. The Cohen’s kappa statistical test reveals that this proposed approach, Cohen’s kappa coefficients are 0.71 for female and 0.66 for male radiographs, p value < 0.05, is in substantial agreement with the bone age assessed by experienced radiologists within the USC dataset. This approach is robust and easy to implement, thus, qualified for computer-aided diagnosis (CAD). The reduced processing time and number of ROIs facilitate BAA.

Assessing the Bone Age of Children in an Automatic Manner Newborn to 18 Years Range

Bone age is assessed through a radiological analysis of the left-hand wrist and is then compared to chronological age. A conflict between these two values indicates an abnormality in the development process of the skeleton. This study, conducted on children aged between 0 and 6 years, proposes a computer-based diagnostic system to eliminate the disadvantages of the methods used in bone age determination. For this purpose, primarily an image processing procedure was applied to the X-ray images of the left-hand wrist of children from different ethnic groups aged between 0 and 6 years. A total of 9 features, corresponding to the carpal bones and distal epiphysis of the radius bone with some physiological attributes of the children, were obtained. Then, by using gain ratio, the best 6 features were used for the classification process. Next, the bone age determination process was performed with the obtained features with the help of the support vector machine (SVM), naive Bayes, k-nearest neighborhood, and C4.5 algorithms. Finally, the features used in the determination process and their effects on the accuracies were examined. The results of the designed system showed that SVM method has a better achievement rate than the other methods at a rate of 72.82%. Additionally, in this study, a new feature corresponding to the distance between the centers of gravity of the carpal bones was used for the classification process, and the analysis of the related feature showed that there was a statistically significant difference at P < 0.05 between this feature and bones in children aged between 0 and 6 years.

/pdf/bone-age-determination-in-young-children-newborn-to-6-years-23im9zzjaf.pdf

Bone age determination in young children (newborn to 6 years old) using support vector machines

Ulke ekonomisi ve kaynaklari bakimindan enerji talebini onceden tahmin etmek cok onemli bir problemdir. Bu calismada, Yercekimi Arama Algoritmasi (YAA) ile YAA’da yapilan bazi yenilikler yapilarak modifiye edilmis ve Modifiye Yercekimi Arama Algoritmasi (MYAA) olarak adlandirilmistir. Enerji talep tahmini, Turkiye’deki ekonomik gostergelerin artisi ile enerji tuketimi arasindaki iliski ile gerceklesmektedir. Enerji talep tahmini icin gayri safi yurtici hasila (GSYH), ithalat, ihracat ve nufus bilgileri hem lineer hem de ussel denklemler kullanilarak tahmin islemi gerceklestirildi. 1997-2011 yillari arasindaki veriler kullanilarak 2017-2037 yillari arasindaki enerji talebi tahmin edilmistir. 2012 ile 2016 yillari ise test verisi olarak kullanilmistir. MGSA ile elde edilen sonuclar GSA sonuclarina gore daha iyi bir tahmin gerceklestirdigi gorulmustur.

/pdf/modified-gravitational-search-algorithm-for-energy-demand-wjpt0u1ea5.pdf

Harun Uğuz

Papers

Bone age determination in young children (newborn to 6 years old) using support vector machines

Modified Gravitational Search Algorithm for Energy Demand Estimation of Turkey

A new hybrid gravitational search-teaching-learning-based optimization method for the solution of economic dispatch of power systems