• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.1-23
• /
• 2018

From the 21st century, various high-quality services have come up with the growth of the internet or 'Information and Communication Technologies'. Especially, the scale of E-commerce industry in which Amazon and E-bay are standing out is exploding in a large way. As E-commerce grows, Customers could get what they want to buy easily while comparing various products because more products have been registered at online shopping malls. However, a problem has arisen with the growth of E-commerce. As too many products have been registered, it has become difficult for customers to search what they really need in the flood of products. When customers search for desired products with a generalized keyword, too many products have come out as a result. On the contrary, few products have been searched if customers type in details of products because concrete product-attributes have been registered rarely. In this situation, recognizing texts in images automatically with a machine can be a solution. Because bulk of product details are written in catalogs as image format, most of product information are not searched with text inputs in the current text-based searching system. It means if information in images can be converted to text format, customers can search products with product-details, which make them shop more conveniently. There are various existing OCR(Optical Character Recognition) programs which can recognize texts in images. But existing OCR programs are hard to be applied to catalog because they have problems in recognizing texts in certain circumstances, like texts are not big enough or fonts are not consistent. Therefore, this research suggests the way to recognize keywords in catalog with the Deep Learning algorithm which is state of the art in image-recognition area from 2010s. Single Shot Multibox Detector(SSD), which is a credited model for object-detection performance, can be used with structures re-designed to take into account the difference of text from object. But there is an issue that SSD model needs a lot of labeled-train data to be trained, because of the characteristic of deep learning algorithms, that it should be trained by supervised-learning. To collect data, we can try labelling location and classification information to texts in catalog manually. But if data are collected manually, many problems would come up. Some keywords would be missed because human can make mistakes while labelling train data. And it becomes too time-consuming to collect train data considering the scale of data needed or costly if a lot of workers are hired to shorten the time. Furthermore, if some specific keywords are needed to be trained, searching images that have the words would be difficult, as well. To solve the data issue, this research developed a program which create train data automatically. This program can make images which have various keywords and pictures like catalog and save location-information of keywords at the same time. With this program, not only data can be collected efficiently, but also the performance of SSD model becomes better. The SSD model recorded 81.99% of recognition rate with 20,000 data created by the program. Moreover, this research had an efficiency test of SSD model according to data differences to analyze what feature of data exert influence upon the performance of recognizing texts in images. As a result, it is figured out that the number of labeled keywords, the addition of overlapped keyword label, the existence of keywords that is not labeled, the spaces among keywords and the differences of background images are related to the performance of SSD model. This test can lead performance improvement of SSD model or other text-recognizing machine based on deep learning algorithm with high-quality data. SSD model which is re-designed to recognize texts in images and the program developed for creating train data are expected to contribute to improvement of searching system in E-commerce. Suppliers can put less time to register keywords for products and customers can search products with product-details which is written on the catalog.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2004.04a
• /
• pp.5-9
• /
• 2004

In this paper, we studied the OCT(Optical Coherence Tomography) system which it has been extensively studied because of having some advantages such as high resolution cross-sectional images, low cost, and small size configuration. A basic principle of OCT system is Michelson interferometer. The characteristics of light source determine the resolution and the transmission depth. As a results, the light source have a commercial SLD with a central wavelength of 1,285 nm and FWHM(Full Width at Half Maximum) of 35.3 nm. The optical delay line part is necessary to equal of the optical path length with scattered light or reflected light from sample. In order to equal the optical path length, the stage which is attached to reference mirror is moved linearly by step motor And the interferometer is configured with the Michelson interferometer using single mod fiber, the scanner can be focused of the sample by using the reference arm. Also, the 2-dimensional cross-sectional images were measured with scanning the transverse direction of the sample by using step motor. After detecting the internal signal of lateral direction at a paint of sample, scanner is moved to obtain the cross-sectional image of 2-demensional by using step motor. Photodiode has been used which has high detection sensitivity, excellent noise characteristic, and dynamic range from 800 nm to 1,700 nm. It is detected mixed small signal between noise and interference signal with high frequency After filtering and amplifying this signal, only envelope curve of interference signal is detected. And then, cross-sectional image is shown through converting this signal into digitalized signal using A／D converter. The resolution of the OCT system is about 30$\mu\textrm{m}$ which corresponds to the theoretical resolution. Also, the cross-sectional image of ping-pong ball is measured. The OCT system is configured with Michelson interferometer which has a low contrast because of reducing the power of feedback interference light. Such a problem is overcomed by using the improved inteferometer. Also, in order to obtain the cross-sectional image within a short time, it is necessary to reduce the measurement time for improving the optical delay line.

• Journal of Broadcast Engineering
• /
• v.13 no.4
• /
• pp.427-439
• /
• 2008

This paper proposes a novel method to convert monoscopic soccer videos to stereoscopic videos. Through the soccer video analysis process, we detect shot boundaries and classify soccer frames into long shot or non-long shot. In the long shot case, the depth mapis generated relying on the size of the extracted ground region. For the non-long shot case, the shot is further partitioned into three types by considering the number of ground blocks and skin blocks which is obtained by a simple skin-color detection method. Then three different depth assignment methods are applied to each non-long shot types: 1) Depth estimation by object region extraction, 2) Foreground estimation by using the skin block and depth value computation by Gaussian function, and 3)the depth map generation for shots not containing the skin blocks. This depth assignment is followed by stereoscopic image generation. Subjective evaluation comparing generated depth maps and corresponding stereoscopic images indicate that the proposed algorithm can yield the sense of depth from a single view images.

• Journal of the Institute of Convergence Signal Processing
• /
• v.8 no.3
• /
• pp.156-162
• /
• 2007

The parallel beam SPECT system acquires projection data by using collimators in conjunction with photon detectors. The projection data of the parallel beam SPECT system is, however, blurred by the point response function of the collimator that is used to define the range of directions where photons can be detected. By increasing the number of parallel holes per unit area in collimator, one can reduce such blurring effect. This approach also, however, has the blurring problem if the distance between the object and the collimator becomes large. In this paper we consider correction methods for artifacts caused by non-circular orbit of parallel beam SPECT with many parallel holes per detector cell. To do so, we model the relationship between the object and its projection data as a linear system, and propose an iterative reconstruction method including artifacts correction. We compute the projector and the backprojector, which are required in iterative method, as a sum of convolutions with distance-dependent point response functions instead of matrix form, where those functions are analytically computed from a single function. By doing so, we dramatically reduce the computation time and memory required for the generation of the projector and the backprojector. We conducted several simulation studies to compare the performance of the proposed method with that of conventional Fourier method. The result shows that the proposed method outperforms Fourier methods objectively and subjectively.

• Journal of the korean academy of Pediatric Dentistry
• /
• v.48 no.3
• /
• pp.245-254
• /
• 2021

The aim of this study was to evaluate the accuracy of 3 different automatic landmark identification programs on lateral cephalgrams and the clinical acceptability in pediatric dentistry. Sixty digital cephalometric radiographs of 7 to 12 years old healthy children were randomly selected. Fourteen landmarks were chosen for assessment and the mean of 3 measurements of each landmark by a single examiner was defined as the baseline landmarks. The mean difference between an automatically identified landmark and the baseline landmark was measured for each landmark on each image. The total mean difference of 3 automatic programs compared to the baseline landmarks were 2.53 ± 1.63 mm. Errors among 3 programs were not significantly different for 12 of 14 landmarks except Orbitale and Gonion. The automatic landmark identification programs showed significant higher mean detection errors than the manual method. The programs couldn't be used as the 1st tool to replace human examiners. But considering short consuming time, these results indicate that all 3 programs have sufficient validity to be used in pediatric dental clinic.

• The Korean Journal of Nuclear Medicine
• /
• v.38 no.2
• /
• pp.180-189
• /
• 2004

Although the outcome of cancer patients after cytotoxic chemotherapy is related diverse mechanisms, multidrug resistance (MDR) for chemotherapeutic drugs due to cellular P-glycoprotein (Pgp) or multidrug-resistance associated protein (MRP) is most important factor in the chemotherapy failure to cancer. A large number of pharmacologic compounds, including verapamil, quinidine, tamoxifen, cyclosporin A and quinolone derivatives have been reported to overcome MDR. Single photon emission computed tomography (SPECT) and positron emission tomography (PET) are available for the detection of Pgp and MRP-mediated transporter. $^{99m}Tc$-MIBI and other $^{99m}Tc$-radiopharmaceuticals are substrates for Pgp and MRP, and have been used in clinical studies for tumor imaging, and to visualize blockade of PgP-mediated transport after modulation of Pgp pump. Colchicine, verapamil and daunorubicin labeled with $^{11}C$ have been evaluated for the quantification of Pgp-mediated transport with PET in vivo and reported to be feasible substrates with which to image Pgp function in tumors. Leukotrienes are specific substrates for MRP and $N-[^{11}C]acetyl-leukotriene$ E4 provides an opportunity to study MRP function non-invasively in vivo. SPECT and PET pharmaceuticals have successfully used to evaluate pharmacologic effects of MDR modulators. Imaging of MDR and reversal of MDR with bioluminescence in a living animal is also evaluated for future clinical trial. We have described recent advances in molecular imaging of MDR and reviewed recent publications regarding feasibility of SPECT and PET imaging to study the functionality of MDR transporters in vivo.

• Journal of radiological science and technology
• /
• v.31 no.3
• /
• pp.287-292
• /
• 2008

Position information of radiation interactions in detection material is essential to reconstruct a radiation source image. With most position sensing techniques, the position information of a single interaction inside the detectors can be precisely obtained. Each interaction position of multi-scattering inside scintillators, however, can not be individually measured and only the average of the scattering positions can be obtained, which causes the uncertainty in the measured interaction position. In this paper, the position uncertainties due to the multi-scattering were calculated by Monte Carlo simulation. The simulation model was a 50 by 50 by 5 mm $LaCl_3$(Ce) scintillator(pixel size is 2 by 2 by 5mm) which was utilized for the dual collimation camera. The dual collimation camera uses the information from both photoelectric effect and Compton scattering, and therefore, position uncertainties for both partial and full energy deposition of radiation interactions are calculated. In the case of partial energy deposition(PED), the standard deviations of positions are less than $1{\sim}2mm$, which means the uncertainty caused by multi-scattering is not significant. Because the effect of the multi-scattering with PED is insignificant, the multi-scattering has little effect on the performance of Compton imaging of dual collimation camera. In the case of full energy deposition(FED), however, the standard deviation of the positions is about twice that of the pixel size of the 1stdetector, except for 122keV incident radiations. Therefore, the standard deviations caused by multi-scatterings should be considered in the design of the coded mask of the dual collimation camera to avoid artifact on the reconstructed image. The position uncertainties of the FEDs are much larger than those of the PEDs for all radiation energies and the ratio of PEDs to FEDs increases when the incident radiation energy increases. The position uncertainties of both PEDs and FEDs are dependent on the incident radiation energy.