• Journal of Convergence for Information Technology
• /
• v.11 no.10
• /
• pp.45-52
• /
• 2021

In this paper, we proposed an Atrous Residual U-Net (AR-UNet) to improve the segmentation accuracy of semantic segmentation method based on U-Net. The U-Net is mainly used in fields such as medical image analysis, autonomous vehicles, and remote sensing images. The conventional U-Net lacks extracted features due to the small number of convolution layers in the encoder part. The extracted features are essential for classifying object categories, and if they are insufficient, it causes a problem of lowering the segmentation accuracy. Therefore, to improve this problem, we proposed the AR-UNet using residual learning and ASPP in the encoder. Residual learning improves feature extraction ability and is effective in preventing feature loss and vanishing gradient problems caused by continuous convolutions. In addition, ASPP enables additional feature extraction without reducing the resolution of the feature map. Experiments verified the effectiveness of the AR-UNet with Cityscapes dataset. The experimental results showed that the AR-UNet showed improved segmentation results compared to the conventional U-Net. In this way, AR-UNet can contribute to the advancement of many applications where accuracy is important.

• BMB Reports
• /
• v.52 no.4
• /
• pp.277-282
• /
• 2019

Currently speaking, it is noted that radiofrequency ablation (RFA) has been the most widely used treatment for hepatocellular carcinoma (HCC) occurring in patients. However, accumulating evidence has demonstrated that the incidence of insufficient RFA (IRFA) may result in the identified rapid progression of residual HCC in the patient, which can greatly hinder the effectiveness and patient reported benefits of utilizing this technique. Although many efforts have been proposed, the underlying mechanisms triggering the rapid progression of residual HCC after IRFA have not yet been fully clarified through current research literature reviews. It was shown in this study that cell proliferation, migration and invasion of residual HepG2 and SMMC7721 cells were significantly increased after the IRFA was simulated in vitro. In other words, it is noted that IRFA could do this by enhancing the image of autophagy of the residual HCC cell via the $HIF-1{\alpha}/BNIP3$ pathway. Consequently, the down-regulation of BNIP3 may result in the inhibition of the residual HCC cell progression and autophagy after IRFA. Our present study results suggest that IRFA could promote residual HCC cell progression in vitro by enhancing autophagy via the $HIF-1{\alpha}/BNIP3$ pathway. For this reason, it is noted that the targeting of the BNIP3 may be useful in preventing the rapid growth and metastasis of residual HCC after IRFA.

• Korean Journal of Metals and Materials
• /
• v.48 no.3
• /
• pp.203-209
• /
• 2010

An application of the instrumented indentation technique has been expanded from the measurements of hardness and elastic modulus to the analysis of residual stress. A slope of the indentation loading curve increases (or decreases) according to compressive (or tensile) residual stress. A theoretical equation has been established for quantifying residual stress from the slope change. However, a precise observation of the remnant indents is indispensible because the theoretical approach needs actual contact information. In addition, the conventional hardness test is still used for predicting the residual stress distribution of welded joints. Thus, we observed the three-dimensional morphologies of the remnant indents formed on artificial stress states and analyzed stress effects on morphological recovery of the indents. First, a depth recovery ratio, which has been regarded as a sensitive stress indicator, did not show a clear dependency with the residual stress. Thus an analysis on volumetric recovery was tried in this study and yielded a inverse proportional behavior with the residual stress. In addition, an elastic to plastic volume recovery ratio showed more significant correlation with the residual stress.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.11
• /
• pp.1653-1660
• /
• 2016

The synthetic aperture rardar (SAR) is an active sensor using microwaves. It transmits a microwave signal, called a chirp pulse, and receives the reflected signal in a moving platform such as satellite and unmanned aerial vehicle. Since this sensor uses microwaves that can penetrate the atmosphere, SAR generates the images regardless of light and weather conditions. However SAR operates on the moving platform, the Doppler shift and the side-looking observation method should be considered. In addtion, a residual image or ghost image can be occurred according to selection of the pulse repetition frequency (PRF). In this paper, a range design of the PRF for the L-band spaceborne SAR system is studied for prevention of SAR image distortion. And the system is studied for prevention of SAR image distortion. And the system parameter and the PRF are calibrated iteratively according to the proposed system design procedure and design constraints. The MATLAB based on SAR system simulator has been developed to verify the validity of calculated PRF. The developed simulator assumes that SAR sensor is operated by the PRF calculated from the design. The results of the simulator show that the targets in image has a valid peak to side-lobe ratio (PSLR) so that the PRF can be used for the spaceborne SAR sensor.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2007.08b
• /
• pp.1688-1691
• /
• 2007

This paper shows that the boundary image sticking can be prohibited completely by using the vacuum sealing process, which means that the residual impurities such as nitrogen or oxygen can be a critical factor inducing the boundary image sticking. The production of boundary image sticking was checked in the test panel fabricated by the $N_2$ or $O_2$ flow during the vacuum sealing process. As a result, the boundary image sticking did not appear in the case of $N_2flow$, whereas the boundary image sticking was observed in the case of $O_2$ flow even though the test panel was fabricated by the vacuum sealing process.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2007.08b
• /
• pp.1478-1481
• /
• 2007

This paper introduced a measurement method for image sticking based on human vision perception. Existing image sticking quantification method is mostly different from visible level by human perception. It takes a long time to measure image sticking which is degraded by time due to using a spot photometer, therefore many test samples could not be evaluated in a given short period of time in mass production line. However, the new measurement method in this paper is possible to evaluate a large quantity of samples in fast and high correlation with human perceptual level of image sticking.

• Journal of Korea Multimedia Society
• /
• v.25 no.5
• /
• pp.676-684
• /
• 2022

As realistic media are widespread in various image processing areas, image or video compression is one of the key technologies to enable real-time applications with limited network bandwidth. Generally, image or video compression cause the unnecessary compression artifacts, such as blocking artifacts and ringing effects. In this study, we propose a Deep Residual Channel-attention Network, so called DRCAN, which consists of an input layer, a feature extractor and an output layer. Experimental results showed that the proposed DRCAN can reduced the total memory size and the inference time by as low as 47% and 59%, respectively. In addition, DRCAN can achieve a better peak signal-to-noise ratio and structural similarity index measure for compressed images compared to the previous methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.4
• /
• pp.1234-1257
• /
• 2023

In this paper, we propose a framework for multi-focus image fusion called PATN. In our approach, by aggregating deep features extracted based on the U-type Transformer mechanism and shallow features extracted using the PSA module, we make PATN feed both long-range image texture information and focus on local detail information of the image. Meanwhile, the edge-preserving information value of the fused image is enhanced using a dense residual block containing the Sobel gradient operator, and three loss functions are introduced to retain more source image texture information. PATN is compared with 17 more advanced MFIF methods on three datasets to verify the effectiveness and robustness of PATN.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2008.10a
• /
• pp.553-556
• /
• 2008

We hereby report novel driving scheme to eliminate effect of "residual" image sticking (RRI) problem which arises due to hysteresis problem in Thin Film Transistor (TFT) in AMOLED Displays. The driving scheme applies "black" voltage after every data voltage period in order to drive AMOLED in uni-direction. The system can be easily implemented with 120 Hz driving scheme which is well matured in AMLCD industries. Our analyses show systematic evaluation of the problem and thereby solving it by simple methods which will be significantly effective of driving OLED towards mass manufacturing stage.

• Journal of Information Processing Systems
• /
• v.14 no.6
• /
• pp.1347-1360
• /
• 2018

Owing to limited energy in wireless devices power saving is very critical to prolong the lifetime of the networks. In this regard, we designed a cross-layer optimization mechanism based on power control in which source node broadcasts a Route Request Packet (RREQ) containing information such as node id, image size, end to end bit error rate (BER) and residual battery energy to its neighbor nodes to initiate a multimedia session. Each intermediate node appends its remaining battery energy, link gain, node id and average noise power to the RREQ packet. Upon receiving the RREQ packets, the sink node finds node disjoint paths and calculates the optimal power vectors for each disjoint path using cross layer optimization algorithm. Sink based cross-layer maximal minimal residual energy (MMRE) algorithm finds the number of image packets that can be sent on each path and sends the Route Reply Packet (RREP) to the source on each disjoint path which contains the information such as optimal power vector, remaining battery energy vector and number of packets that can be sent on the path by the source. Simulation results indicate that considerable energy saving can be accomplished with the proposed cross layer power control algorithm.