• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.1
• /
• pp.55-57
• /
• 2015

This paper presents an adaptive non-local means denoising algorithm using down-scaled images. This work provides a method to reduce artifacts and information loss around context region by increasing the number of similar patches for high activity region with down-scaled images. Experimental results demonstrate that the proposed algorithm outperforms the non-local means algorithm more than 1.5 (dB).

• Journal of Biomedical Engineering Research
• /
• v.26 no.4
• /
• pp.185-191
• /
• 2005

In this paper, we proposed a downscaled 3D object technique using medical images for telediagnostic use. The proposed system consisted of downscaling/thresholding processes for building a downscaled 3D object and a process for obtaining 2D images at specific angles for diagnosis support. We used 80 slices of Digital Imaging and Communication in Medicine(DICOM) CT images as sample images and the platform-independent Java language for the experiment. We confirmed that the total image set size and transmission time of the original DICOM image set using a down-scaled 3D object decreased approximately $99\%\;and\;98.41\%,$ respectively. With additional studies, the proposed technique obtained from these results will become useful in supporting diagnosis for home and hospital care.

• Journal of the Korea Computer Graphics Society
• /
• v.17 no.1
• /
• pp.39-44
• /
• 2011

In this paper, we introduce a novel method to reduce image to a small size, such that the quality of the image is improved when it is up-scaled. Recent hardwares including cameras and display devices allow us to capture and display high-resolution images. However, it is not always realistic to store and transmit those high-resolution images due to limitation of storage and network bandwidth. Therefore, high-resolution images are often down-scaled to be stored and transmitted, and then up-scaled back for display. To improve final image quality in this scenario, we first formulate selected up-scale methods as linear transformations. The optimal reduction methods are obtained as its inverse transformation. Based on this basic idea, we develop down-scale kernel that is optimized for each up-scale method. In our experiment, the proposed method could improve the quality of the up-scaled image noticeable.

• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.107-108
• /
• 2007

In this paper, a cosine-modulated filter bank with a modified synthesis part is proposed for arbitrary scaling of images, whereby down/up-sampling kernels based on a compactly supported sampling function (CSSF) are utilized. Also, an optimized adaptive interpolation technique is incorporated into the filter bank structure to compensate for quality degradation arising in scaled images. Finally, simulation results verify that high quality images with arbitrary sizes can be obtained by applying the proposed approach.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.7
• /
• pp.2328-2344
• /
• 2022

2D human pose estimation still faces difficulty in low-resolution images. Most existing top-down approaches scale up the target human bonding box images to the large size and insert the scaled image into the network. Due to up-sampling, artifacts occur in the low-resolution target images, and the degraded images adversely affect the accurate estimation of the joint positions. To address this issue, we propose a multi-resolution input feature fusion network for human pose estimation. Specifically, the bounding box image of the target human is rescaled to multiple input images of various sizes, and the features extracted from the multiple images are fused in the network. Moreover, we introduce a guiding channel which induces the multi-resolution input features to alternatively affect the network according to the resolution of the target image. We conduct experiments on MS COCO dataset which is a representative dataset for 2D human pose estimation, where our method achieves superior performance compared to the strong baseline HRNet and the previous state-of-the-art methods.

• Journal of the Korea Computer Graphics Society
• /
• v.8 no.3
• /
• pp.9-16
• /
• 2002

Recently, high dynamic range (HDR) compression has attracted much attention due to the wide availability of HDR images. In this paper, we present an HDR compression method using a progressive image, which is a multi-level image representation based on a progressive mesh. An HDR image can be decomposed into a base image and a sequence of details by conversion into a progressive image. This decomposition provides a good structure to highly compress the dynamic range while preserving image details. The base image and larger details are considerably scaled down but smaller details are slightly scaled down. Experimental results show that our method successfully generates HDR compressed images without halo artifacts by controlling two intuitive parameters.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.27 no.2
• /
• pp.131-137
• /
• 2016

In this paper, a feasibility on a ground-penetrating radar for detecting subsurface cavities near buried pipes has been investigated. The experimental setup was implemented by employing an impulse ground-penetrating radar system, a xy Cartesian coordinate robot, an underground material filled tank, a metal pipe and a simulated cavity model. In particular, the simulated cavity model was constructed by packing Styrofoam chips and balls, which have both similar electrical properties to an air-filled cavity and a solid shape. Through typical three experiments, B-scan data of the radar have been acquired and displayed as 2-D gray-scale images. According to the comparison of B-scan images, we show that the subsurface cavities near the buried pipes can be detected by using the radar survey.

• Smart Structures and Systems
• /
• v.32 no.3
• /
• pp.135-151
• /
• 2023

Using 3D asphalt pavement surface data, a deep and multiscale network named CrackNet-M is proposed in this paper for pixel-level crack detection for improvements in both accuracy and robustness. The CrackNet-M consists of four function-specific architectural modules: a central branch net (CBN), a crack map enhancement (CME) module, three pooling feature pyramids (PFP), and an output layer. The CBN maintains crack boundaries using no pooling reductions throughout all convolutional layers. The CME applies a pooling layer to enhance potential thin cracks for better continuity, consuming no data loss and attenuation when working jointly with CBN. The PFP modules implement direct down-sampling and pyramidal up-sampling with multiscale contexts specifically for the detection of thick cracks and exclusion of non-crack patterns. Finally, the output layer is optimized with a skip layer supervision technique proposed to further improve the network performance. Compared with traditional supervisions, the skip layer supervision brings about not only significant performance gains with respect to both accuracy and robustness but a faster convergence rate. CrackNet-M was trained on a total of 2,500 pixel-wise annotated 3D pavement images and finely scaled with another 200 images with full considerations on accuracy and efficiency. CrackNet-M can potentially achieve crack detection in real-time with a processing speed of 40 ms/image. The experimental results on 500 testing images demonstrate that CrackNet-M can effectively detect both thick and thin cracks from various pavement surfaces with a high level of Precision (94.28%), Recall (93.89%), and F-measure (94.04%). In addition, the proposed CrackNet-M compares favorably to other well-developed networks with respect to the detection of thin cracks as well as the removal of shoulder drop-offs.

• Journal of computational fluids engineering
• /
• v.20 no.2
• /
• pp.96-102
• /
• 2015

In this study, the green water phenomena on rectangular shaped structure is numerically simulated by STAR-CCM+ to investigate the flow pattern including the velocity profiles in bubbly water flow. 5 phases of the formation of green water in front of and over the rectangular shaped structure is simulated at the design condition which is scaled down by 1:125 from FPSO operating in GOM. All numerical results are compared with the experimental results performed in a two dimensional wave flume. The water deformation due to the green water are obtained by the high speed CCD camera with employing the shadow graphy technique, which is allowed to take the bubbly water flow into images. A series of image taken by shadow graphy technique is analyzed with MQD method to calculate the velocity in bubbly water flow.

• Journal of the Korean Society of Costume
• /
• v.59 no.9
• /
• pp.131-145
• /
• 2009

Literature in beauty art is insufficient regarding beauty spot, though it has played an important role in beauty art history. This study aims at analyzing various patterns and decorative functions of beauty spots and examining their remarkable differences by place and time. It attempts to find the symbolistic characteristics of beauty patch which influenced European countries socially and culturally in the 17th and 18th centuries, findings are as follows: first, beauty spots are extremely various by patch patterns and by forms, e.g. Wha-Jeon(花鈿), Bindi and artificial points. In the view of cosmetic purpose, they have been used to beautify appearances, reveal social positions and express the desires like blessing or protection from evil. In the times of Baroque and Rococo, though temporarily, people with beauty spots and their belonged societies used them as communicative tools: obviously, they functioned symbolistically. This study surveyed 156 persons to find what cosmetic messages and images beauty spots present in the contemporary world. Using SPSS 13.0, the study analyzed the responses and found that beauty spots are still functioning and welcomed to let women look more beautiful and more attractive, though their cosmetic messages and sizes have been simplified and scaled down.