• Journal of Periodontal and Implant Science
• /
• v.48 no.2
• /
• pp.114-123
• /
• 2018

Purpose: The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Methods: Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. Results: The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. Conclusions: We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

• Journal of Korea Multimedia Society
• /
• v.16 no.7
• /
• pp.810-819
• /
• 2013

FCM (Fuzzy C-Means) clustering algorithm, a typical split-based clustering algorithm, has been successfully applied to the various fields. Nonetheless, the FCM clustering algorithm has some problems, such as high sensitivity to noise and local data, the different clustering result from the intuitive grasp, and the setting of initial round and the number of clusters. To address these problems, in this paper, we determine fuzzy numbers which project the FCM clustering result on the axis with the specific attribute. And we propose a model that the fuzzy numbers apply to FDT (Fuzzy Decision Tree). This model improves the two problems of FCM clustering algorithm such as elevated sensitivity to data, and the difference of the clustering result from the intuitional decision. And also, this paper compares the effect of the proposed model and the result of FCM clustering algorithm through the experiment using real traffic and rainfall data. The experimental results indicate that the proposed model provides more reliable results by the sensitivity relief for data. And we can see that it has improved on the concordance of FCM clustering result with the intuitive expectation.

• Journal of Korean Society of Transportation
• /
• v.20 no.2
• /
• pp.135-148
• /
• 2002

In this paper, a queue length calculation algorithm using image detectors has been proposed. The algorithm produces the queue length using a pair of image detectors installed both on upstream and on downstream of a corridor. In addition, a new framework for controlling the traffic signal system based on queue length has been presented. More specifically, the scheme of determining the cycle time and green split using the queue lengths has been proposed. To validate the results, a simulation study was conducted with a network environment. Results showed that the proposed method gave better operational performance than a traditional method. However, additional validation effort is necessary in order to apply the real traffic conditions.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.4E
• /
• pp.52-60
• /
• 2001

This paper describes real time implementation of subband sound localization system on a floating-point DSP TI TMS320C31. The system determines two dimensional location of an active speaker in a closed room environment with real noise presents. The system consists of an two microphone array connected to TI DSP hosted by PC. The implemented sound localization algorithm is Subband CPSP which is an improved version of traditional CPSP (Cross-Power Spectrum Phase) method. The algorithm first split the input speech signal into arbitrary number of subband using subband filter banks and calculate the CPSP in each subband. It then averages out the CPSP results on each subband and compute a source location estimate. The proposed algorithm has an advantage over CPSP such that it minimize the overall estimation error in source location by limiting the specific band dominant noise to that subband. As a result, it makes possible to set up a robust real time sound localization system. For real time simulation, the input speech is captured using two microphone and digitized by the DSP at sampling rate 8192 hz, 16 bit/sample. The source location is then estimated at once per second to satisfy real-time computational constraints. The performance of the proposed system is confirmed by several real time simulation of the speech at a distance of 1m, 2m, 3m with various speech source locations and it shows over 5% accuracy improvement for the source location estimation.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.34 no.8
• /
• pp.47-54
• /
• 2006

In this paper, numerical algorithms applied in two-step eulerian scheme are investigated and implemented. Element quantities are advected with donor or van Leer algorithm. Nodal quantities are advected with Simplifed ALE [SALE] algorithm. Material interfaces are determined with Simple Line Interface Calculation [SLIC] algorithm. Practical aspects considered for code development are addressed in detail. The results of developed two-step Eulerian code are verified by comparing with those from pure lagrangian scheme and commercial code.

• Journal of information and communication convergence engineering
• /
• v.4 no.3
• /
• pp.123-129
• /
• 2006

The limited number of roads and the increasing number of vehicles demand the automatic regulation of overspeed vehicles, illegal vehicles, and overloaded vehicles and the automatic charge calculation depending on the type of the vehicle. To meet such requirements, it is important to remove the shadow of the vehicle as processing and recognizing an image captured by a camera. The shadow of the vehicle is likely to cause misclassification of the vehicle type due to diverse errors and mistakes occurring when detecting geometrical properties of the vehicle. In case that shadows of two different vehicles are overlapped, not only the type of the vehicles may be misclassified but also it is difficult to accurately identify the type of the vehicles. In this paper, we propose a robust algorithm to remove the shadow of a vehicle by calculating the luminance, the chrominance, the gradient density of the cast shadow from information acquired using the image subtraction of the background, and to recognize the substantial vehicle figure. Even when it is hard to detect and split a target vehicle from its shadow as shadows of vehicles are attached to each other, our robust algorithm can detect the vehicle figure only. We implemented our system with a general camera and conducted experiments on various vehicles on general roads to find out our vehicle shade removal algorithm is efficient when detecting and recognizing vehicles.

• Earthquakes and Structures
• /
• v.27 no.3
• /
• pp.177-189
• /
• 2024

Accurately predicting the failure modes of reinforced concrete (RC) columns is essential for structural design and assessment. In this study, the challenges of imbalanced datasets and complex feature selection in machine learning (ML) methods were addressed through an optimized ML approach. By combining feature selection and oversampling techniques, the prediction of seismic failure modes in rectangular RC columns was improved. Two feature selection methods were used to identify six input parameters. To tackle class imbalance, the Borderline-SMOTE1 algorithm was employed, enhancing the learning capabilities of the models for minority classes. Eight ML algorithms were trained and fine-tuned using k-fold shuffle split cross-validation and grid search. The results showed that the artificial neural network model achieved 96.77% accuracy, while k-nearest neighbor, support vector machine, and random forest models each achieved 95.16% accuracy. The balanced dataset led to significant improvements, particularly in predicting the flexure-shear failure mode, with accuracy increasing by 6%, recall by 8%, and F1 scores by 7%. The use of the Borderline-SMOTE1 algorithm significantly improved the recognition of samples at failure mode boundaries, enhancing the classification performance of models like k-nearest neighbor and decision tree, which are highly sensitive to data distribution and decision boundaries. This method effectively addressed class imbalance and selected relevant features without requiring complex simulations like traditional methods, proving applicable for discerning failure modes in various concrete members under seismic action.

• Journal of Digital Convergence
• /
• v.12 no.11
• /
• pp.379-386
• /
• 2014

In this paper, the processing speed of Grabcut algorithm in order to efficiently improve the GPU (Graphics Processing Unit) for processing the data from the method. Grabcut algorithm has excellent performance object detection algorithm. Grabcut existing algorithms to split the foreground area and the background area, and then background and foreground K-cluster is assigned a cluster. And assigned to gradually improve the results, until the process is repeated. But Drawback of Grabcut algorithm is the time consumption caused by the repetition of clustering. Thus GPGPU (General-Purpose computing on Graphics Processing Unit) using the repeated operations in parallel by processing Grabcut algorithm to effectively improve the processing speed of the method. We proposed method of execution time of the algorithm reduced the average of about 95.58%.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.3
• /
• pp.59-71
• /
• 2018

This study was initiated to develop an optimal signal control algorithm for coordinated signalized intersections using individual vehicle's information which can be collected in a format of prove vehicle data (PVD) via V2I (Vehicle to Infrastructure) communication environment. For developing this signal optimization algorithm, three modules were developed for phase group length computation, split distribution, and phase sequence assignment. The simulation analysis using the microscopic simulation model, Vissim, was conducted for evaluating the effectiveness of the developed algorithm. The analysis result represented that the performance of the developed algorithm is far superior to that of the fixed coordinated signal control method which is the most common signal control method for coordinated signalized intersections in Korea.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.8
• /
• pp.1155-1160
• /
• 2013

The rapid growth of smart phones has made location-based services (LBSs) widely available. However, the use of LBS can raise privacy issues, as LBS can allow adversaries to violate the location privacy of users. There has been a considerable amount of research on preserving user location privacy. Most of these studies try to preserve location privacy by achieving what is known as location K-anonymity. In this paper, we propose a hierarchical clustering-based spatial cloaking algorithm for LBSs. The proposed algorithm constructs a tree using a modified version of agglomerative hierarchical clustering. The experimental results show, in terms of the ASR size, that the proposed algorithm is better than Hilbert Cloak and comparable to RC-AR (R-tree Cloak implementation of Reciprocal with an Asymmetric R-tree split). In terms of the ASR generation time, the proposed algorithm is much better in its performance than RC-AR and similar in performance to Hilbert Cloak.