• Journal of Biomedical Engineering Research
• /
• v.24 no.2
• /
• pp.55-60
• /
• 2003

For quantitative analysis of the cardiac diseases. it is necessary to segment the left-ventricle (LY) in MR (Magnetic Resonance) cardiac images. Snake or active contour model has been used to segment LV boundary. However, the contour of the LV front these models may not converge to the desirable one because the contour may fall into local minimum value due to image artifact inside of the LY Therefore, in this paper, we Propose the Preprocessing method using k-means clustering and merging algorithms that can improve the performance of the active contour model. We verified that our proposed algorithm overcomes local minimum convergence problem by experiment results.

• Journal of Intelligence and Information Systems
• /
• v.11 no.1
• /
• pp.95-105
• /
• 2005

As the security system using facial recognition, the recognition performance depends on the environments (e. g. face expression, hair style, age and make-up etc.) For the revision of easily changeable environment, it's generally used to set up the threshold, replace the face image which covers the threshold into images already registered, and update the face images additionally. However, this usage has the weakness of inaccuracy matching results or can easily active by analogous face images. So, we propose the genetic algorithm which absorbs greatly the facial similarity degree and the recognition target variety, and has excellence studying capacity to avoid registering inaccuracy. We experimented variable and similar face images (each 30 face images per one, total 300 images) and performed inherent face images based on ingredient analysis as face recognition technique. The proposed method resulted in not only the recognition improvement of a dominant gene but also decreasing the reaction rate to a recessive gene.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.2
• /
• pp.358-363
• /
• 2009

Text region detection from the natural scene images used in a variety of applications, many research are needed in this field. Recent research methods is to detect the text region using various algorithm which it is combination of edge based and connected component based. Therefore, this paper proposes an text region detection using edge and regional minima/maxima transformation algorithm from natural scene images, and then detect the connected components of edge and regional minima/maxima, labeling edge and regional minima/maxima connected components. Analysis the labeled regions and then detect a text candidate regions, each of detected text candidates combined and create a single text candidate image, Final text region validated by comparing the similarity and adjacency of individual characters, and then as the final text regions are detected. As the results of experiments, proposed algorithm improved the correctness of text regions detection using combined edge and regional minima/maxima connected components detection methods.

• Journal of the Korea Convergence Society
• /
• v.9 no.7
• /
• pp.1-7
• /
• 2018

Feature selection, one of data preprocessing techniques, is one of major research areas in many applications dealing with large dataset. It has been used in pattern recognition, machine learning and data mining, and is now widely applied in a variety of fields such as text classification, image retrieval, intrusion detection and genome analysis. The proposed method is based on a genetic algorithm which is one of meta-heuristic algorithms. There are two methods of finding feature subsets: a filter method and a wrapper method. In this study, we use a wrapper method, which evaluates feature subsets using a real classifier, to find an optimal feature subset. The training dataset used in the experiment has a severe class imbalance and it is difficult to improve classification performance for rare classes. After preprocessing the training dataset with SMOTE, we select features and evaluate them with various machine learning algorithms.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.2
• /
• pp.55-60
• /
• 2011

In this paper, we propose a new method for extracting muscles from lumbar images. The proposed method sets areas without distortions with field expert's assistance as areas of measuring interest and removing noises from initial ultrasonic videos. Then, the method emphasizes the brightness contrast with Ends-in search stretching algorithm and separate thoracic vertebra from subcutaneous fat area using morphological characteristics. 4-directions contour tracing algorithm is applied to extract the bottom of subcutaneous fat area. Extracting thoracic vertebra area requires noise removal and morphological characteristics as well among candidate areas obtained by controlling min-max brightness. The thickness of muscles is then defined as the length between subcutaneous fat area and extracted thoracic vertebra. The experiment which consists of 368 image analysis verifies that the proposed method is more effective in measuring the thickness of muscles than before.

• Korean Institute of Interior Design Journal
• /
• v.24 no.2
• /
• pp.32-41
• /
• 2015

We live in the era of unlimited design competition. As the importance of design is increasing in all areas including marketing, each country does its best effort on design development. However, the preparation on protecting interior design rights by intellectual property laws(IPLs) has not been enough even though they occupy an important place in the design field. It is not quite easy to make a judgement on the similarity between two images having a single common factor because the factors which are composed of interior design have complicated interactive relations between them. From the IPLs point of view, designs with the similar overall appearance are decided to be similar. Objective evaluation criteria not only for designers but also for design examiners and judges are required in order to protect interior design by the IPLs. The objective of this study is the analysis of the possibility that a computer algorithm method can be useful to decide the similarity of interior design images. According to this study, it is realized that the Img2 which is one of content-based image retrieval computer programs can be utilized to measure the degree of the similarity. The simulation results of three descriptors(CEDD, FCTH, JCD) in the Img2 showed the high degree of similar patterns compared with the results of perceptual judgment by observers. In particular, it was verified that the Img2 has high availability on interior design images with a high score of similarity below 60 which are perceptually judged by observers.

• Atmosphere
• /
• v.27 no.2
• /
• pp.119-131
• /
• 2017

We present and discuss the Tropopause Folding Turbulence Detection (TFTD) algorithm for the Korean Communication, Ocean, Meteorological Satellite (COMS) which is originally developed for the Tropopause Folding Turbulence Product (TFTP) from the Geostationary Operational Environmental Satellite (GOES)-R. The TFTD algorithm assumes that the tropopause folding is linked to the Clear Air Turbulence (CAT), and thereby the tropopause folding areas are detected from the rapid spatial gradients of the upper tropospheric specific humidity. The Layer Averaged Specific Humidity (LASH) is used to represent the upper tropospheric specific humidity calculated using COMS $6.7{\mu}m$ water vapor channel and ERA-interim reanalysis temperature at 300, 400, and 500 hPa. The comparison of LASH with the numerical model specific humidity shows a strong negative correlation of 80% or more. We apply the single threshold, which is determined from sensitivity analysis, for cloud-clearing to overcome strong gradient of LASH at the edge of clouds. The tropopause break lines are detected from the location of strong LASH-gradient using the Canny edge detection based on the image processing technique. The tropopause folding area is defined by expanding the break lines by 2-degree positive gradient direction. The validations of COMS TFTD is performed with Pilot Reports (PIREPs) filtered out Convective Induced Turbulence (CIT) from Dec 2013 to Nov 2014 over the South Korea. The score test shows 0.49 PODy (Probability of Detection 'Yes') and 0.64 PODn (Probability of Detection 'No'). Low POD results from various kinds of CAT reported from PIREPs and the characteristics of high sensitivity in edge detection algorithm.

• Journal of Magnetics
• /
• v.21 no.2
• /
• pp.286-292
• /
• 2016

This study evaluates the change of computer tomography (CT) number in the case of the metal artifact reduction (MAR) algorithm, using the phantom. The images were obtained from dual CT using a gammex 467 tissue characterization phantom, which is similar to human tissues. The test method was performed by dividing pre and post MAR algorithm and measured CT values of nonmagnetic materials within the phantom. In addition, the changes of CT values for each material were compared and analyzed after measuring CT values up to 140 keV, using the spectral HU curve followed by CT scan. As a result, in the cases of N rod (trabecular bone) and E rod (trabecular bone), the CT numbers decreased as keV increasing but were constant above 90 keV. In the cases of I rod (dense bone) and K rod (dense bone), the CT numbers also decreased as keV increased but were uniform above 90 keV. The CT numbers from 40 keV to 140 keV were consistent in the cases of J rod (liver), D rod (liver), L rod (muscle), and F rod (muscle). For A rod (adipose), G rod (adipose), B rod (breast) and O rod (breast), the CT numbers increased as keV increased but were constant after 90 keV. The CT numbers from 40 keV to 140 keV were consistent in the cases of C rod (lung (exhale)), P rod (lung (exhale)), M rod (lung (inhale)) and H rod (lung (exhale)). Conclusively, because dual CT exhibits no changes in image quality and is able to analyze nonmagnetic materials by measuring the CT values of various materials, it will be used in the future as a useful tool for the diagnosis of lesions.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.6
• /
• pp.464-470
• /
• 2016

In this paper, we propose an improved face detection algorithm and determination method for drowsiness status of driver from the opening and closing frequency of the detected eye. For this purpose, face, eyes, nose, and mouth are detected based on conventional Viola-Jones face detection algorithm and spatial correlation of face. Here the spatial correlation of face is performed by DFP(Detect Face Part) based on seven characteristics. The experimental results on Caltect face image database revealed that the detection rates of noise particularly showed the improved performance of 13.78% in comparison to that of the previous Viola-Jones algorithm. Furthermore, we analyze the driver's drowsiness determination cumulative value of the eye closed state as a function of time based on SVM (Support Vector Machine) and PERCLOS(Percentage Closure of Eyes). The experimental results confirmed the usefulness of the proposed method by obtaining a driver's drowsiness determination rate of 93.28%.

• Journal of the Korea Computer Graphics Society
• /
• v.16 no.1
• /
• pp.1-7
• /
• 2010

With the widespread use of vision-based surveillance systems, the capability for person identification is now an essential component. However, the CCTV cameras used in surveillance systems tend to produce relatively low-resolution images, making it difficult to use face recognition techniques for person identification. Therefore, an algorithm is proposed for person identification in CCTV camera images based on the clothing. Whenever a person is authenticated at the main entrance of a building, the clothing feature of that person is extracted and added to the database. Using a given image, the clothing area is detected using background subtraction and skin color detection techniques. The clothing feature vector is then composed of textural and color features of the clothing region, where the textural feature is extracted based on a local edge histogram, while the color feature is extracted using octree-based quantization of a color map. When given a query image, the person can then be identified by finding the most similar clothing feature from the database, where the Euclidean distance is used as the similarity measure. Experimental results show an 80% success rate for person identification with the proposed algorithm, and only a 43% success rate when using face recognition.