• Korean Journal of Remote Sensing
• /
• v.39 no.6_1
• /
• pp.1211-1224
• /
• 2023

Clouds cause many difficult problems in observing land surface phenomena using optical satellites, such as national land observation, disaster response, and change detection. In addition, the presence of clouds affects not only the image processing stage but also the final data quality, so it is necessary to identify and remove them. Therefore, in this study, we developed a new cloud detection technique that automatically performs a series of processes to search and extract the pixels closest to the spectral pattern of clouds in satellite images, select the optimal threshold, and produce a cloud mask based on the threshold. The cloud detection technique largely consists of three steps. In the first step, the process of converting the Digital Number (DN) unit image into top-of-atmosphere reflectance units was performed. In the second step, preprocessing such as Hue-Value-Saturation (HSV) transformation, triangle thresholding, and maximum likelihood classification was applied using the top of the atmosphere reflectance image, and the threshold for generating the initial cloud mask was determined for each image. In the third post-processing step, the noise included in the initial cloud mask created was removed and the cloud boundaries and interior were improved. As experimental data for cloud detection, CAS500-1 L2G images acquired in the Korean Peninsula from April to November, which show the diversity of spatial and seasonal distribution of clouds, were used. To verify the performance of the proposed method, the results generated by a simple thresholding method were compared. As a result of the experiment, compared to the existing method, the proposed method was able to detect clouds more accurately by considering the radiometric characteristics of each image through the preprocessing process. In addition, the results showed that the influence of bright objects (panel roofs, concrete roads, sand, etc.) other than cloud objects was minimized. The proposed method showed more than 30% improved results(F1-score) compared to the existing method but showed limitations in certain images containing snow.

• Journal of the Korean Society of Radiology
• /
• v.5 no.2
• /
• pp.59-66
• /
• 2011

There is a difficulty because of classifying the anatomical structure in the neck CT scan by the beam hardening artifact no more than disease and it including the 6, 7 number cervical spine and intervertebral disk. In case of enforcing the neck CT scan cause of the inner diameter of beam artifact tried to be inquired by the image evaluation according to the change of the image evaluation according to the direction of the shoulder joint applying the variation method of a posture and location and Kernel value and it was most appropriate, the lion tax and Kernel value try to be searched for through an experiment. Somatom Sensation 16 (Siemens, Enlarge, Germany) equipment was used in a patient 30 people coming to the hospital for the neck CT scan. A workstation used the AW 4.4 version (GE, USA). According to a direction and location of the shoulder joint, the patient posture gave a change to the direction of the shoulder joint as the group S it gave a change as three postures and placed the both arms comfortably and helps a group N and augmented unipolar left in the wealthy merchant and group P it memorized the both hands and ordered the eversion and drops below to the utmost and enforced a scan. By using a reconstructing method as the second opinion, it gave and reconstructed the Kernel value a change based on scan data with B 10 (very smooth), B 20 (smooth), B 30 (medium smooth), B 40 (medium), B 50 (medium sharp), B 60 (sharp), and B 70 (very sharp). By using image data which gave the change of the examination posture and change of the Kernel value and are obtained, we analyzed through the noise value measurement and image evaluation of. The outside wire eversion orders the both hands and the examination posture is cost in the neck CT scan with the group P it drops below to the utmost. And in case of when reconstructing with B 40 (medium) or B 50 (medium sharp) being most analyzed into the inappropriate posture and Kernel value and applying the Kernel value to a clinical, it is considered to be very useful.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.37 no.5
• /
• pp.84-93
• /
• 2000

The conventional fractal decoding was implemented to IFS(iterated function system) for every range regions But a part of the range regions can be decoded without the iteration and there is a data dependence regions In order to decode $R{\times}R$ range blocks, It needs $2R{\times}2R$ domain blocks This decoding can be analyzed to the dependence graph The vertex of the graph represents the range blocks, and the vertex is classified into the vertex of the range and domain The edge indicates that the vertex is referred to the other vertices The in-degree and the out-degree are defined to the number of the edge that is entered and exited, respectively The proposed method is analyzed by a dependence graph to the fractal code, and the decoding order is recomposed by the information of the out-degree That is, If the out-degree of the vertex is zero, then this vertex can be used to the vertex with data dependence Thus, the proposed method can extend the data dependence regions by the recomposition of the decoding order As a result, the Iterated regions are minimized without loss of the image quality or PSNR(peak signal-to-noise ratio), Therefore, it can be a fast decoding by the reducing to the computational complexity for IFS in the fractal Image decoding.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.19 no.1
• /
• pp.30-36
• /
• 2015

Purpose Whole body bone scan, which makes up a largest percentage of nuclear medicine tests, has high sensitivity and resolution about bone lesion like osteomyelitis, fracture and the early detection of primary cancer. However, any standard for valuation has not yet been created except minimum factor. Therefore, in this study, we will analysis the method which show a quantitative evaluation index in whole body bone scan. Materials and Methods This study is conducted among 30 call patients, who visited the hospital from April to September 2014 with no special point of view about bone lesion, using GE INFINIA equipment. Enumerated data is measured mainly with patient's whole body count and lumbar vertabrae, and the things which include CNR (Contrast to Noise ratio), SNR (Signal to Noise ratio) are calculated according to the mean value signal and standard deviation of each lumbar vertabrae. In addition, the numerical value with the abdominal thickness is compared to each value by the change of scan speed and tissue equivalent material throughout the phantom examination, and compared with 1hours deleyed value. Completely, on the scale of ten, 2 reading doctors and 5 skilled radiologists with 5-years experience analysis the correlation between visual analysis with blind test and quantitative calculation. Results The whole body count and interest region count of patients have no significant correlation with visual analysis value throughout the blind test(P<0.05). There is definite correlation among CNR and SNR. In phantom examination, Value of the change was caused by the thickness of the abdomen and the scan speed. And The poor value of the image in the subject as a delay test patient could be confirmed that the increase tendency. Conclusion Now, a standard for valuation has not been created in whole body bone scan except minimum factor. In this study, we can verify the significant correlation with blind test using CNR and SNR and also assure that the scan speed is a important factor to influence the imagine quality from the value. It is possible to be some limit depending on the physiology function and fluid intake of patient even if we progress the evaluation in same condition include same injection amount, same scan speed and so on. However, that we prove the significant evaluation index by presenting quantitative calculation objectively could be considered academic value.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.47 no.4
• /
• pp.30-39
• /
• 2010

Since Diffusion tensor from Diffusion Tensor Magnetic Resonance Imaging(DT-MRI) is so sensitive to noise, the principle eigenvector(PEV) calculated from Diffusion tensor could be erroneous. Tractography obtained from PEV could be deviated from the real fiber tract. Therefore regularization process is needed to eliminate noise. In this paper, to reduce noise in DT-MRI measurements, the Dyadic Sorting(DS) method as regularization of the eigenvalue and the eigenvector is applied in the tractography. To resort the eigenvalues and the eignevectors, the DS method uses the intervoxel overlap function which can measure the overlap between eigenvalue-eigenvector pairs in the $3\times3$ pixel. In this paper, we applied the DS method to the three-dimensional volume. We discuss the error analysis and numerical study to the synthetic and the experimental data. As a result, we have shown that the DS method is more efficient than the median filtering methods as much as 79.97%~83.64%, 85.62%~87.76% in AAE, AFA respectively for the corticospinal tract of the experimental data.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.18 no.1
• /
• pp.69-75
• /
• 2014

Purpose: The aim of this study was to evaluate changes of quantitative and semi-quantitative myocardial perfusion indices and image quality by image reconstruction methods in $^{13}N$-ammonia ($^{13}N-NH_3$) myocardial perfusion PET/CT. Materials and Methods: Data of 14 (8 men, 6 women) patients underwent rest and adenosine stress $^{13}N-NH_3$ PET/CT (Biograph TruePoint 40 with TrueV, Siemens) were collected. Listmode scans were acquired for 10 minutes by injecting 370MBq of $^{13}N-NH_3$. Dynamic and static reconstruction was performed by use of FBP, iterative2D (2D), iterative3D (3D) and iterative TrueX (TrueX) algorithm. Coronary flow reserve (CFR) of dynamic reconstruction data, extent(%) and total perfusion deficit (TPD) (%) measured in sum of 4-10 minutes scan were evaluated by comparing with 2D method which was recommended by vendor. The image quality of each reconstructed data was compared and evaluated by five nuclear medicine physicians through a blind test. Results: CFR were lower in TrueX 18.68% (P=0.0002), FBP 4.35% (P=0.1243) and higher in 3D 7.91% (P<0.0001). As semi-quantitative values, extent and TPD of stress were higher in 3D 3.07%p (P=0.001), 2.36%p (P=0.0002), FBP 1.93%p (P=0.4275), 1.57%p (P=0.4595), TrueX 5.43%p (P=0.0003), 3.93%p (P<0.0001). Extent and TPD of rest were lower in FBP 0.86%p (P=0.1953), 0.57%p (P=0.2053) and higher in 3D 3.21%p (P=0.0006), 2.57%p (P=0.0001) and TrueX 5.36%p (P<0.0001), 4.36%p (P<0.0001). Based on the results of the blind test for image resolution and noise from the snapshot, 3D obtained the highest score, followed by 2D, TrueX and FBP. Conclusion: We found that quantitative and semi-quantitative myocardial perfusion values could be under- or over-estimated according to the reconstruction algorithm in $^{13}N-NH_3$ PET/CT. Therefore, proper dynamic and static reconstruction method should be established to provide accurate myocardial perfusion value.

• Journal of radiological science and technology
• /
• v.44 no.2
• /
• pp.85-90
• /
• 2021

The purpose of this study is to consider usefulness of using AEC mode and importance of patient center location in L-spine lateral radiography by comparing dose and image quality according to the change of patient center location with using AEC mode or not. In this study, guide wire is attached to the human body phantom's lumbar spine and the lead ruler is attached to the bottom of the wall detector to find out center location in detector. ESD, mAs, and EI were selected as dose factors, and image quality was compared through SNR. With the lumbar spine located center of the detector, dose factors and image quality were compared according to using AEC mode or not. Afterwards, phantom moved 4 cm and 8 cm back and forth and compared dose factors and image quality. The exposure parameters were 85 kVp, 320 mA, x-ray field size 10×17 inch, and the distance between the center X-ray and the detector was fixed at 100 cm. The center X-ray was perpendicular to the fourth lumbar spine and the only bottom AEC chamber was used. All data were analyzed by independent t-test and ANOVA. As a result of this study, with AEC when the center is matched, ESD was 1.31±0.01 mGy, without AEC was 2.12±0.01 mGy. SNR was shown to be 22.81±1.83, and 23.44±1.87 respectively. When the phantom's center moves 4 cm, 8 cm forward, and 4 cm, 8 cm backward, ESD were 1.09±0.004 mGy, 0.32±0.003 mGy, 1.19±0.017 mGy, 1.11±0.006 mGy respectively, SNR were 18.29±0.60 dB, 11.11±0.22 dB, 18.98±0.80 dB, 17.71±0.82 dB. Using AEC in L-spine lateral radiography reduced ESD by 38%, EI by 35%, and mAs by 38%, without any difference in SNR(p<0.05). When the phantom's center moves 4 cm, 8 cm forward, and 4 cm, 8 cm backward, ESD was decreasing each 16%, 75%, 9%, 15%, EI was decreasing each 14%, 77%, 15%, 20%, mAs was decreasing each 15% 75% 9%, 15%. SNR was decreasing each 19%, 51%, 17%, 22%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.10
• /
• pp.3955-3971
• /
• 2020

Due to the significant increase in the use of autonomous car technology, it is essential to integrate this technology with high-precision digital map data containing more precise and accurate roadway information, as compared to existing conventional map resources, to ensure the safety of self-driving operations. While existing map technologies may assist vehicles in identifying their locations via Global Positioning System, it is however difficult to update the environmental changes of roadways in these maps. Roadway vision algorithms can be useful for building autonomous vehicles that can avoid accidents and detect real-time location changes. We incorporate a hybrid architectural design that combines unsupervised classification of vision data with supervised joint fusion classification to achieve a better noise-resistant algorithm. We identify, via a deep learning approach, an intelligent hybrid fusion algorithm for fusing multimodal vision feature data for roadway classifications and characterize its improvement in accuracy over unsupervised identifications using image processing and supervised vision classifiers. We analyzed over 93,000 vision frame data collected from a test vehicle in real roadways. The performance indicators of the proposed hybrid fusion algorithm are successfully evaluated for the generation of roadway digital maps for autonomous vehicles, with a recall of 0.94, precision of 0.96, and accuracy of 0.92.

• Proceedings of the KIEE Conference
• /
• 2003.11b
• /
• pp.295-298
• /
• 2003

In this paper, a novel maximum a posterion (MAP) estimation for the channel decoding of H.264 codes in the presence of transmission error is presented. Arithmetic codes with a forbidden symbol and trellis search techniques are employed in order to estimate the best transmitted. And, there has been growing interest of communication, the research about transmission of exact data is increasing. Unlike the case of voice transmission, noise has a fatal effect on the image transmission. The reason is that video coding standards have used the variable length coding. So, only one bit error affects the all video data compressed before resynchronization. For reasons of that, channel needs the channel codec, which is robust to channel error. But, usual channel decoder corrects the error only by channel error probability. So, designing source codec and channel codec, Instead of separating them, it is tried to combine them jointly. And many researches used the information of source redundancy In received data. But, these methods do not match to the video coding standards, because video ceding standards use not only one symbol but also many symbols in same data sequence. In this thesis, We try to design combined source-channel codec that is compatible with video coding standards. This MAP decoder is proposed by adding semantic structure and semantic constraint of video coding standards to the method using redundancy of the MAP decoders proposed previously. Then, We get the better performance than usual channel coder's.

• Korean Journal of Digital Imaging in Medicine
• /
• v.10 no.1
• /
• pp.45-50
• /
• 2008

Recently, MR Cholangiography used mainly bu controlling of patient's breathing. There is breathing hold techniques to get images within shopt time and gating technique adjusted to respiration cycle for high resolution image. In this study, the aim of this experiment is to know on clinical usefulness compared with PACE and RTG thchniques. This study's period is from 2006 in November to 2007 in January. A total of 21 patients investigated at MAGNETOM Sonata 1.5T (SIEMENS Erlangen) with use of 12ch body coil. MR acquisition protocol used 3D turbo spin echo coronal sequence. Scan parameters applied to potimal setting in use as gating techniques, respectively. Analysis of consuming timing evaluated with rapidness. As analysis of quantity, the common bile duct, gall bladder measured in signal intensities, then these data were calculated by signal to noise ratio and contrast to noise ratio. Qualitative analysis, experienced 2radiologists and 3 RTs were evaluated into 3groups about artifact, accuracy of lesions, sharpness of the common bile duct or gall bladder. As a result of analysis, when compared to PACE, consuming time of the RTG took less than PACE, On both CNRs and SNRs, PACE technique was slightly high values than RTG(p<0.05). Qualitative analysis' results, discrimination of lesions in the common bile duct, gall bladder get a significance level in both RTG and PACE techniques but presence's artifact of breathing and pulsation highly demonstrate in PACE techniques. In conclusion, both PACE and RTG methods at MRCP provided prominently clinical information for the common bile duct, gall bladder. If machines have not limitation with performance, induction of breathing holding also will help getting diagnistic quality.