• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2000.11b
• /
• pp.270-278
• /
• 2000

A machine vision system using charge coupled device(CCD) camera for the weed detection in a radish farm was developed. Shape features were analyzed with the binary images obtained from color images of radish and weeds. Aspect, Elongation and PTB were selected as significant variables for discriminant models using the STEPDISC option. The selected variables were used in the DISCRIM procedure to compute a discriminant function for classifying images into one of the two classes. Using discriminant analysis, the successful recognition rate was 92% for radish and 98% for weeds. To recognize radish and weeds more effectively than the discriminant analysis, an artificial neural network(ANN) was used. The developed ANN model distinguished the radish from the weeds with 100%. The performance of ANNs was improved to prevent overfitting and to generalize well using a regularization method. The successful recognition rate in the farms was 93.3% for radish and 93.8% for weeds. As a whole, the machine vision system using CCD camera with the artificial neural network was useful to detect weeds in the radish farms.

• Korean Journal of Radiology
• /
• v.20 no.4
• /
• pp.662-670
• /
• 2019

Objective: A developmental venous anomaly (DVA) is a vascular malformation of ambiguous clinical significance. We aimed to quantify the susceptibility of draining veins (χ_vein) in DVA and determine its significance with respect to oxygen metabolism using quantitative susceptibility mapping (QSM). Materials and Methods: Brain magnetic resonance imaging of 27 consecutive patients with incidentally detected DVAs were retrospectively reviewed. Based on the presence of abnormal hyperintensity on T2-weighted images (T2WI) in the brain parenchyma adjacent to DVA, the patients were grouped into edema (E+, n = 9) and non-edema (E-, n = 18) groups. A 3T MR scanner was used to obtain fully flow-compensated gradient echo images for susceptibility-weighted imaging with source images used for QSM processing. The χ_vein was measured semi-automatically using QSM. The normalized χ_vein was also estimated. Clinical and MR measurements were compared between the E+ and E- groups using Student's t-test or Mann-Whitney U test. Correlations between the χ_vein and area of hyperintensity on T2WI and between χ_vein and diameter of the collecting veins were assessed. The correlation coefficient was also calculated using normalized veins. Results: The DVAs of the E+ group had significantly higher χ_vein (196.5 ± 27.9 vs. 167.7 ± 33.6, p = 0.036) and larger diameter of the draining veins (p = 0.006), and patients were older (p = 0.006) than those in the E- group. The χ_vein was also linearly correlated with the hyperintense area on T2WI (r = 0.633, 95% confidence interval 0.333-0.817, p < 0.001). Conclusion: DVAs with abnormal hyperintensity on T2WI have higher susceptibility values for draining veins, indicating an increased oxygen extraction fraction that might be associated with venous congestion.

• Imaging Science in Dentistry
• /
• v.42 no.3
• /
• pp.183-190
• /
• 2012

Purpose: The purpose of this study was to investigate the level of clinical image quality of panoramic radiographs and to analyze the parameters that influence the overall image quality. Materials and Methods: Korean dental clinics were asked to provide three randomly selected panoramic radiographs. An oral and maxillofacial radiology specialist evaluated those images using our self-developed Clinical Image Quality Evaluation Chart. Three evaluators classified the overall image quality of the panoramic radiographs and evaluated the causes of imaging errors. Results: A total of 297 panoramic radiographs were collected from 99 dental hospitals and clinics. The mean of the scores according to the Clinical Image Quality Evaluation Chart was 79.9. In the classification of the overall image quality, 17 images were deemed 'optimal for obtaining diagnostic information,' 153 were 'adequate for diagnosis,' 109 were 'poor but diagnosable,' and nine were 'unrecognizable and too poor for diagnosis'. The results of the analysis of the causes of the errors in all the images are as follows: 139 errors in the positioning, 135 in the processing, 50 from the radiographic unit, and 13 due to anatomic abnormality. Conclusion: Panoramic radiographs taken at local dental clinics generally have a normal or higher-level image quality. Principal factors affecting image quality were positioning of the patient and image density, sharpness, and contrast. Therefore, when images are taken, the patient position should be adjusted with great care. Also, standardizing objective criteria of image density, sharpness, and contrast is required to evaluate image quality effectively.

• Imaging Science in Dentistry
• /
• v.49 no.4
• /
• pp.295-299
• /
• 2019

Purpose: This clinical pilot study was performed to determine the effectiveness of dual-energy cone-beam computed tomography (DE-CBCT) in measuring bone mineral density (BMD). Materials and Methods: The BMD values obtained using DE-CBCT were compared to those obtained using calibrated multislice computed tomography (MSCT). After BMD calibration with specially designed phantoms, both DE-CBCT and MSCT scanning were performed in 15 adult dental patients. Three-dimensional (3D) Digital Imaging and Communications in Medicine data were imported into a dental software program, and the defined regions of interest (ROIs) on the 3-dimensional surface-rendered images were identified. The automatically-measured BMD values of the ROIs (g/㎤), the differences in the measured BMD values of the matched ROIs obtained by DE-CBCT and MSCT 3D images, and the correlation between the BMD values obtained by the 2 devices were statistically analyzed. Results: The mean BMD values of the ROIs for the 15 patients as assessed using DE-CBCT and MSCT were 1.09±0.07 g/㎤ and 1.13±0.08 g/㎤, respectively. The mean of the differences between the BMD values of the matched ROIs as assessed using DE-CBCT and calibrated MSCT images was 0.04±0.02 g/㎤. The Pearson correlation coefficient between the BMD values of DE-CBCT and MSCT images was 0.982 (r=0.982, P<0.001). Conclusion: The newly developed DE-CBCT technique could be used to measure jaw BMD in dentistry and may soon replace MSCT, which is expensive and requires special facilities.

• ETRI Journal
• /
• v.28 no.4
• /
• pp.521-524
• /
• 2006

In this letter, we propose a novel computational integral imaging reconstruction technique based on a lenslet array model. The proposed technique provides improvement of viewing images by extracting multiple pixels from elemental images according to ray tracing based on the lenslet array model. To show the feasibility of the proposed technique, we analyze it according to ray optics and present the experimental results.

• Journal of Information Display
• /
• v.12 no.1
• /
• pp.37-41
• /
• 2011

Presented herein is a novel stereoscopic three-dimensional (3D) display system with active color filter glasses. This system provides full-color 3D images by applying the time-multiplexing technique on the original anaglyph method. By switching between the opposite anaglyph statuses, a full-color anaglyph is presented. A liquid crystal panel from a 3D monitor serves as an active color filter operating at 120 Hz. A display panel and a color filter are connected to one graphic card as a dual-link system, for synchronization. To test the quality of this system, a left/right-eye image separation test and an experiment with stereoscopic images were carried out. Although there was some crosstalk and blur, the system, as expected, provided full-color 3D display. This system overcomes a monochromatic 3D image, which is the major weakness of the original anaglyph system.

• Investigative Magnetic Resonance Imaging
• /
• v.25 no.3
• /
• pp.156-163
• /
• 2021

Recently, neuromelanin and nigrosome imaging techniques have been developed to evaluate the substantia nigra in Parkinson's disease. Previous studies have shown potential benefits of quantitative analysis of neuromelanin and nigrosome images in the substantia nigra, although visual assessments have been performed to evaluate structures in most studies. In this study, we investigate the potential of using deep learning based automatic region segmentation techniques for quantitative analysis of the substantia nigra. The deep convolutional neural network was trained to automatically segment substantia nigra regions on 3D nigrosome and neuromelanin sensitive MR images obtained from 30 subjects. With a 5-fold cross-validation, the mean calculated dice similarity coefficient between manual and deep learning was 0.70 ± 0.11. Although calculated dice similarity coefficients were relatively low due to empirically drawn margins, selected slices were overlapped for more than two slices of all subjects. Our results demonstrate that deep convolutional neural network-based method could provide reliable localization of substantia nigra regions on neuromelanin and nigrosome sensitive MR images.

• Korean Journal of Remote Sensing
• /
• v.21 no.3
• /
• pp.243-252
• /
• 2005

Wavelet transform and its inverse processing provide the effective framework for data fusion. The purpose of this study is to investigate applicability of wavelet transform using texture images for the urban remote sensing application. We tried several experiments regarding image fusion by wavelet transform and texture imaging using high resolution images such as IKONOS and KOMPSAT EOC. As for texture images, we used homogeneity and ASM (Angular Second Moment) images according that these two types of texture images reveal detailed information of complex features of urban environment well. To find out the useful combination scheme for further applications, we performed DWT(Discrete Wavelet Transform) and IDWT(Inverse Discrete Wavelet Transform) using texture images and original images, with adding edge information on the fused images to display texture-wavelet information within edge boundaries. The edge images were obtained by the LoG (Laplacian of Gaussian) processing of original image. As the qualitative result by the visual interpretation of these experiments, the resultant image by each fusion scheme will be utilized to extract unique details of surface characterization on urban features around edge boundaries.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.1
• /
• pp.25-29
• /
• 2023

In this paper, the purpose is to create a standard of AI training dataset type for commercial space design. As the market size of the field of space design continues to increase and the time spent increases indoors after COVID-19, interest in space is expanding throughout society. In addition, more and more consumers are getting used to the digital environment. Therefore, If you identify trends and preemptively propose the atmosphere and specifications that customers require quickly and easily, you can increase customer trust and conduct effective sales. As for the data set type, commercial districts were divided into a total of 8 categories, and images that could be processed were derived by refining 4,009,30MB JPG format images collected through web crawling. Then, by performing bounding and labeling operations, we developed a 'Dataset for AI Training' of 3,356 commercial space image data in CSV format with a size of 2.08MB. Through this study, elements of spatial images such as place type, space classification, and furniture can be extracted and used when developing AI algorithms, and it is expected that images requested by clients can be easily and quickly collected through spatial image input information.

• Journal of Biomedical Engineering Research
• /
• v.28 no.4
• /
• pp.549-555
• /
• 2007

In order to study cortical properties in human, it is necessary to obtain an accurate and explicit representation of the cortical surface in individual subjects. Among many approaches, surface-based method that reconstructs a 3-D model from contour lines on cross-section images is widely used. In general, however, medical brain imaging has some problems such as the complexity of the images, non-linear gain artifacts and so on. Due these limitations, therefore, extracting anatomical structures from imaging data is very a complicated and time-consuming task. In this paper, we present an improved method for extracting contour lines of cortical surface from magnetic resonance images that simplifies procedures of a conventional method. The conventional method obtains contour lines through thinning and chain code process. On the other hand, the proposed method can extract contour lines from comparison between boundary data and labeling image without supplementary processes. The usefulness of the proposed method has been verified using brain image.