• Journal of the Korean Society of Radiology
• /
• 제15권4호
• /
• pp.547-554
• /
• 2021

In order to quickly and accurately diagnose pneumonia on a chest X-ray image, different batch sizes of 4, 8, 16, and 32 were applied to the same Xception deep learning model, and modeling was performed 3 times, respectively. As a result of the performance evaluation of deep learning modeling, in the case of modeling to which batch size 32 was applied, the results of accuracy, loss function value, mean square error, and learning time per epoch showed the best results. And in the accuracy evaluation of the Test Metric, the modeling applied with batch size 8 showed the best results, and the precision evaluation showed excellent results in all batch sizes. In the recall evaluation, modeling applied with batch size 16 showed the best results, and for F1-score, modeling applied with batch size 16 showed the best results. And the AUC score evaluation was the same for all batch sizes. Based on these results, deep learning modeling with batch size 32 showed high accuracy, stable artificial neural network learning, and excellent speed. It is thought that accurate and rapid lesion detection will be possible if a batch size of 32 is applied in an automatic diagnosis study for feature extraction and classification of pneumonia in chest X-ray images using deep learning in the future.

• Food Science of Animal Resources
• /
• 제33권5호
• /
• pp.640-645
• /
• 2013

The purpose of this study was to utilize the results as a basic data of benzoic acids in animal products that didn't mention in the quality standard of National Veterinary Research and Quarantine Service (NVRQS) to solve the conflict of international trade and administration. Set-Pak method listed in the quality standard of NVRQS, faster than auto distillation methods with same recovery selected as a pre treatment for the determination of benzoic acid. The regression curve of benzoic acid with Sep-Pak method was linear with the $R^2$ value of 0.999 and the limit of detection (LOD) and limit of quantitation (LOQ) was 0.058 mg/kg and 0.176 mg/kg, respectively. The benzoic acid in the fermented milk was detected after the fermentation stage by addition of starter culture with the level of 2.28~10.48 mg/kg and 0~16.5 mg/kg in the commercial fermented milk products without detection by the addition of syrup. In case of cheese products, the benzoic acids level was influenced by the curd formation (Camembert cheese) and the quality of natural cheese (processed cheese), by the way, the benzoic acid level of commercial natural cheese was 0~4.2 mg/kg, processed cheese was 0~20.8 mg/kg, respectively. Based on this result, it may be possible to utilize as a basic data for the systematic control the level of natural benzoic acids in raw material, processing and final products of animal origin.

• Economic and Environmental Geology
• /
• 제55권5호
• /
• pp.551-561
• /
• 2022

Geophysical exploration methods are very useful for generating high-resolution images of underground structures, and such methods can be applied to investigation of buried cultural properties and for determining their exact locations. In this study, image feature extraction and image segmentation methods were applied to automatically distinguish the structures of buried relics from the high-resolution ground-penetrating radar (GPR) images obtained at the center of Silla Kingdom, Gyeongju, South Korea. The major purpose for image feature extraction analyses is identifying the circular features from building remains and the linear features from ancient roads and fences. Feature extraction is implemented by applying the Canny edge detection and Hough transform algorithms. We applied the Hough transforms to the edge image resulted from the Canny algorithm in order to determine the locations the target features. However, the Hough transform requires different parameter settings for each survey sector. As for image segmentation, we applied the connected element labeling algorithm and object-based image analysis using Orfeo Toolbox (OTB) in QGIS. The connected components labeled image shows the signals associated with the target buried relics are effectively connected and labeled. However, we often find multiple labels are assigned to a single structure on the given GPR data. Object-based image analysis was conducted by using a Large-Scale Mean-Shift (LSMS) image segmentation. In this analysis, a vector layer containing pixel values for each segmented polygon was estimated first and then used to build a train-validation dataset by assigning the polygons to one class associated with the buried relics and another class for the background field. With the Random Forest Classifier, we find that the polygons on the LSMS image segmentation layer can be successfully classified into the polygons of the buried relics and those of the background. Thus, we propose that these automatic classification methods applied to the GPR images of buried cultural heritage in this study can be useful to obtain consistent analyses results for planning excavation processes.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• 제7권1호
• /
• pp.129-135
• /
• 2003

Optical fiber has many advantages, such as high reliability, long lifetime, immunity to the electromagnetic interference, high speed response and low cost. In this study, we proposed and developed an optical fiber impact and pressure sensor for prevention of accident which occurs in the automatic system or auto door. The principle of the sensor is to detect different optical intensity caused by variation of a speckle pattern due to the external perturbation. Speckle pattern appears at the end of a multimode fiber in which coherent beam propagates. The fabricated sensor in this study was tested. As a result of experiments, amplitude of the output signal isn't linear, but it has sufficient sensitivity for a sensor. Moreover, we can control sensitivity of the sensor by an amplifier at receiver. It has several advantages which are ability of detection at all point on the multimode fiber, large sensitive area, and many application areas for a sensing impact and pressure.

• Proceedings of the KSR Conference
• /
• 한국철도학회 2011년도 정기총회 및 추계학술대회 논문집
• /
• pp.1966-1972
• /
• 2011

Optical flow estimation based on multi constraint approaches is frequently used for recognition of moving objects. This paper proposed the method to monitor passenger boarding using image processing when a train is operated based on Automatic Train Operation(ATO). The movement of passenger can be detected to compare two images, one is a basic image and another is immediately captured by CCTV. Optical Flow helps to find the movement of passenger when two images are compared. The movement of passenger is one of important informations for ATO system because it needs to decide door status.

• Journal of KIISE:Software and Applications
• /
• 제37권9호
• /
• pp.706-711
• /
• 2010

In the applications of a human speech interface, reducing the error rate in recognition is the one of the main research issues. Many previous studies attempted to correct errors using post-processing, which is dependent on a manually constructed corpus and correction patterns. We propose an automatically learnable post-processing method that is independent of the characteristics of both the domain and the speech recognizer. We divide the entire post-processing task into two steps: error detection and error correction. We consider the error detection step as a classification problem for which we apply the conditional random fields (CRFs) classifier. Furthermore, we apply transformation-based learning (TBL) to the error correction step. Our experimental results indicate that the proposed method corrects a speech recognizer's insertion, deletion, and substitution errors by 25.85%, 3.57%, and 7.42%, respectively.

• Journal of KIISE:Software and Applications
• /
• 제33권10호
• /
• pp.851-861
• /
• 2006

The paper proposes a novel unsupervised video object segmentation algorithm for image sequences with low depth-of-field (DOF), which is a popular photographic technique enabling to represent the intention of photographer by giving a clear focus only on an object-of-interest (OOI). The proposed algorithm largely consists of two modules. The first module automatically extracts OOIs from the first frame by separating sharply focused OOIs from other out-of-focused foreground or background objects. The second module tracks OOIs for the rest of the video sequence, aimed at running the system in real-time, or at least, semi-real-time. The experimental results indicate that the proposed algorithm provides an effective tool, which can be a basis of applications, such as video analysis for virtual reality, immersive video system, photo-realistic video scene generation and video indexing systems.

• Aerospace Engineering and Technology
• /
• 제3권1호
• /
• pp.242-250
• /
• 2004

This paper describes the detection method of wire-like obstacles using millimeter-wave radar system. Passive sensor like CCD camera can be used for the detection of high power electric cables on the hills or mountains and it can give very good quality of obstacle target information. But this system is very limited to use by bad weather condition. The detection capability for different diameters of wire targets using millimeter radar system have been accomplished. To simulate the target on the moving helicopter, rotating targets are used with fixed radar system. In the experiment 11mm, 16mm and 22mm diameter of wires have been detected in single, two and three wires in one position. The detected signal from single wire was very clear on gray level image. Three wires placed very closely together could be recognized in range, cross range image plane. For two and three wires, blur effect due to mutual scattering effect is observed.

• The Journal of Society for e-Business Studies
• /
• 제19권4호
• /
• pp.45-60
• /
• 2014

Dentist can obtain 3D anatomical information without distortion and information loss by using dental Computed Tomography scan images on line, and also can make the preoperative plan of implant placement or orthodontics. It is essential to segment individual tooth for making an accurate diagnosis. However, it is very difficult to distinguish the difference in the brightness between the dental and adjacent area. Especially, the root of a tooth is very elusive to automatically identify in dental CT images because jawbone normally adjoins the tooth. In the paper, we propose a method of automatically tooth region segmentation, which can identify the root of a tooth clearly. This algorithm separate the tooth from dental CT scan images by using Seeded Region Growing method on dental crown and by using Level-set method on dental root respectively. By using the proposed method, the results can be acquired average 19.2% better accuracy, compared to the result of the previous methods.

• Korean Journal of Construction Engineering and Management
• /
• 제23권2호
• /
• pp.45-53
• /
• 2022

The purpose of this study is to present a model for analyzing construction errors of reinforced concrete pile foundations using drones. First, a drone is used to obtain an aerial image of the construction site, and an orthomosaic image is generated based on those images. Then, the circular pile foundation is automatically recognized from the orthomosaic image by using the Hough transform circle detection method. Finally, the distance is calculated based on the the center point of the reinforced concrete pile foundation in the overlapped data. As a case study, the proposed concrete concrete pile foundation construction quality control model was applied to the real construction site in Incheon to evaluate the proposed model.