• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.1
• /
• pp.82.3-82.3
• /
• 2019

In this study, we generate future full disk magnetograms in 12, 24, 36 and 48 hours advance from SDO/HMI images using deep learning. To perform this generation, we apply the convolutional generative adversarial network (cGAN) algorithm to a series of SDO/HMI magnetograms. We use SDO/HMI data from 2011 to 2016 for training four models. The models make AI-generated images for 2017 HMI data and compare them with the actual HMI magnetograms for evaluation. The AI-generated images by each model are very similar to the actual images. The average correlation coefficient between the two images for about 600 data sets are about 0.85 for four models. We are examining hundreds of active regions for more detail comparison. In the future we will use pix2pix HD and video2video translation networks for image prediction.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.35 no.3
• /
• pp.195-202
• /
• 2017

Ortho images created by aerial photogrammetry have been used in large areas but they are uneconomical for small areas and continuous change observation. The drones have been developed for military purposes, and recently they are being used crop management and analysis, broadcast relay, meteorological observation and disaster investigation and so on. Also there were a lot of studies of expensive commercial drone. In this paper, lower price self-assembly drone usable for in small areas, Obtained images and produced Orthomosaic and DSM using mission planner which is a normal digital camera and open source program, and postprocessing was used Pix4d software. GCP errors are X-coordinate 3.4cm, Y-coordinate 2.4cm, Z-coordinate 4.2cm. It seems like the self-assembly drone can be used for various fields.

• Journal of The Korean Astronomical Society
• /
• v.37 no.4
• /
• pp.185-191
• /
• 2004

We have been developing a solar observing system based on a fast CCD camera 1M30P made by the DALSA company. Here we examine and present the characteristics and performance of the camera. For this we have analyzed a number of images of a flat wall illuminated by a constant light source. As a result we found that in the default operating mode 1) the mean bias level is 49 ADU/pix, 2) the mean dark current is about 8 ADU /s/pix, 3) the readout noise is 1.3 ADU, and 4) the gain is about 42 electrons/ ADU. The CCD detector is found to have a linearity with a deviation smaller than $6\%$, and a uniform sensitivity better than $1\%$. These parameters will be used as basic inputs in the analysis of data to be taken by the camera.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.350-352
• /
• 2022

발병률이 낮은 병은 데이터 불균형 문제가 발생하며, 이는 의료계에서 겪는 원초적인 문제이다. 이런 불균형 문제를 해결하고자 Pix2Pix 로 생성적 적대 신경망 기반 의료 이미지 증강 기법을 설계하여 데이터 불균형 문제 해결 및 성능을 향상시켰다. 합성 데이터의 추가 및 기하학적 데이터 증강의 유무에 대한 4 가지 시나리오로 성능을 비교하여 제안된 기법이 가장 효과적임을 보인다.

• Journal of Broadcast Engineering
• /
• v.27 no.1
• /
• pp.69-79
• /
• 2022

In this study, a method of generating and utilizing a normal map image used to represent the characteristics of the surface of an image material to improve the classification accuracy of the original material image is proposed. First of all, (1) to generate a normal map that reflects the surface properties of a material in an image, a U-Net with attention-R2 gate as a generator was used, and a Pix2Pix-based method using the generated normal map and the similarity with the original normal map as a reconstruction loss was used. Next, (2) we propose a network that can improve the accuracy of classification of the original material image by applying the previously created normal map image to the attention gate of the classification network. For normal maps generated using Pixar Dataset, the similarity between normal maps corresponding to ground truth is evaluated. In this case, the results of reconstruction loss function applied differently according to the similarity metrics are compared. In addition, for evaluation of material image classification, it was confirmed that the proposed method based on MINC-2500 and FMD datasets and comparative experiments in previous studies could be more accurately distinguished. The method proposed in this paper is expected to be the basis for various image processing and network construction that can identify substances within an image.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.24 no.1
• /
• pp.15-19
• /
• 2020

Purpose Generative Adversarial Network(GAN) is one of deep learning technologies. This is a way to create a real fake image after learning the real image. In this study, after acquiring artificial intelligence images through GAN, We were compared and evaluated with real scan time images. We want to see if these technologies are potentially useful. Materials and Methods 30 patients who underwent ¹⁸F-FDG Brain PET/CT scanning at Severance Hospital, were acquired in 15-minute List mode and reconstructed into 1,2,3,4,5 and 15minute images, respectively. 25 out of 30 patients were used as learning images for learning of GAN and 5 patients used as verification images for confirming the learning model. The program was implemented using the Python and Tensorflow frameworks. After learning using the Pix2Pix model of GAN technology, this learning model generated artificial intelligence images. The artificial intelligence image generated in this way were evaluated as Mean Square Error(MSE), Peak Signal to Noise Ratio(PSNR), and Structural Similarity Index(SSIM) with real scan time image. Results The trained model was evaluated with the verification image. As a result, The 15-minute image created by the 5-minute image rather than 1-minute after the start of the scan showed a smaller MSE, and the PSNR and SSIM increased. Conclusion Through this study, it was confirmed that AI imaging technology is applicable. In the future, if these artificial intelligence imaging technologies are applied to nuclear medicine imaging, it will be possible to acquire images even with a short scan time, which can be expected to reduce artifacts caused by patient movement and increase the efficiency of the scanning room.

• Journal of Plant Biotechnology
• /
• v.30 no.2
• /
• pp.135-141
• /
• 2003

Peroxiredoxin(Pix) are large family of peroxidases that reduce alkyl hydroperoxides and hydrogen peroxide. A cDNA clone (referred to as swPrxl) encoding Pix was from a sweetpotato cDNA library constructed from suspension-sultured cells, and its expression was investigated in terms of stress. The swPrxl contained an open reading frame (ORF) encoding mature protein of 193 amino acids with calculated molecular mass of 20.8kDa. The predicted amino acid sequence of swPrxl has two conserved cysteines that are essential resicues for the reduction of peroxides. It showed high amino acid sequence homology ot PixIIF of Arabidopsis (77%) and putative Prx of rice(72%). RNA gel-blot analysis showed that swPrxl gene was expressed dominantly in leave among intact tissues, and also highly detect in suspension-cultured cells. Interestingly, the level of swPrxl transcripts was almost the same regardless of the growth stage in suspension culture. Furthermore, the transcription level of swPrxl gene was not significantly changed in response to various stress treatments such as wounding, extreme temperature and stress-related chemicals RT-PCR analyses.

• Journal of Cadastre & Land InformatiX
• /
• v.50 no.2
• /
• pp.155-168
• /
• 2020

Open source is a key growth engine of the 4th industrial revolution, and the continuous development and use of various algorithms for image processing is expected. The purpose of this study is to examine the effectiveness of the UAS image processing open source based algorithm by comparing and analyzing the water reproduction and moving object filtering function and the time required for data processing in 3D reproduction. Five matching algorithms were compared based on recall and processing speed through the 'ANN-Benchmarks' program, and HNSW (Hierarchical Navigable Small World) matching algorithm was judged to be the best. Based on this, 108 algorithms for image processing were constructed by combining each methods of triangulation, point cloud data densification, and surface generation. In addition, the 3D reproduction and data processing time of 108 algorithms for image processing were studied for UAS (Unmanned Aerial System) images of a park adjacent to the sea, and compared and analyzed with the commercial image processing software 'Pix4D Mapper'. As a result of the study, the algorithms that are good in terms of reproducing water and filtering functions of moving objects during 3D reproduction were specified, respectively, and the algorithm with the lowest required time was selected, and the effectiveness of the algorithm was verified by comparing it with the result of 'Pix4D Mapper'.

• Proceedings of the Microbiological Society of Korea Conference
• /
• 2004.05a
• /
• pp.141-146
• /
• 2004

SPIN90 was identified to farm molecular complex with $\betaPIX$, WASP and Nck. This complex shows that SPIN90 interacts with Nck in a manner dependent upon cell adhesion to extracellular matrix, but $SPIN90{\cdot}{\beta}PIX{\cdot}WASP$ complex was stable even in suspended cells. This suggests that SPIN90 serves as an adaptor molecule to recruit other proteins to Nck at focal adhesions. SPIN90 was phosphorylated by ERK1, which was, itself, activated by cell adhesion and platelet-derived growth factor. Such phosphorylation of SPIN90 likely promotes the interaction of the $SPIN90{\cdot}{\beta}PIX{\cdot}WASP$ complex and Nck. It thus appears that the interaction of the $SPIN90{\cdot}{\beta}PIX{\cdot}WASP$ complex with Nck is crucial for stable cell adhesion and can be dynamically modulated by SPIN90 phosphorylation that is dependent on cell adhesion and ERX activation. SPIN90 directly binds syndapin I, syndapin isoform II-1 and II-s via its PRD region in vitro, in vivo and also associates with endocytosis core components such as clathrin and dynamin. In neuron and fibroblast, SPIN90 colocalizes with syndapins as puntate form, consistent with a role for SPIN90 in clathrin-mediated endocytosis pathway. Overexpression of SPIN90 N-term inhibits receptor-mediated endocytosis. Interestingly, SPIN90 PRD, binding interface of syndapin, significantly blocks internalization of transferrin, demonstrating SPIN90 involvement in endocytosis in vivo by interacting syndapin. Depletion of endogenous SPIN90 by introducing $\alpha-SPIN90$ also blocks receptor-mediated endocytosis. Actin polymerization could generate farce facilitating the pinch-out event in endocytosis, detach newly formed endocytic vesicle from the plasma membrane or push out them via the cytosol on actin tails. Here we found that SPIN90 localizes to high actin turn over cortical area, actin-membrane interface and membrane ruffle in PDGF treated cells. Overexpression of SPIN90 has an effect on cortical actin rearrangement as filopodia induction and it is mediated by the Arp2/3 complex at cell periphery. Consistent with a role in actin organization, CFP-SPIN90 present in actin comet tail generated by PIP5 $kinase\gamma$ overexpression. Therefore this study suggests that SPIN90 is functional linker between endocytosis and actin cytoskeleton.

• Journal of Pharmaceutical Investigation
• /
• v.35 no.5
• /
• pp.361-367
• /
• 2005

In this study, we evaluated and compared the pharmacodynamics of paclitaxel (PTX) in human A549 NSCLC cells grown as monolayers or as three-dimensional histocultures. Growth inhibitory effects were determined after incubating cells in drug free medium until 96 hr post drug exposure initiation. Cell cycle arrest and apoptosis were measured by flow cytometry. The growth inhibition induced by PTX was significantly different in monolayers and histocultures, and PTX showed significantly less cytotoxicity in histocultures where large resistant fractions were observed. Moreover, although PIX induced significant $G_{2}/M$ arrest followed by apoptosis in monolayers in a drug concentration-dependant manner, $G_{2}/M$ arrest was not elicited in histocultures. However, apoptotic cells appeared from the $G_{2}/M$ phase in histocultures. In this study, we provide first evidence that PIX in three-dimensional histocultures, does not induce $G_{2}/M$ arrest, but rather that it induces $G_{2}/M$ phase specific apoptosis. Overall, our data demonstrate different pharmacodynamics of PTX in traditional monolayer and three-dimensional histocultures.