• ETRI Journal
• /
• v.46 no.3
• /
• pp.513-525
• /
• 2024

Deep neural networks trained on labeled medical data face major challenges owing to the economic costs of data acquisition through expensive medical imaging devices, expert labor for data annotation, and large datasets to achieve optimal model performance. The heterogeneity of diseases, such as Alzheimer's disease, further complicates deep learning because the test cases may substantially differ from the training data, possibly increasing the rate of false positives. We propose a reconstruction-based self-supervised anomaly detection model to overcome these challenges. It has a dual-subnetwork encoder that enhances feature encoding augmented by skip connections to the decoder for improving the gradient flow. The novel encoder captures local and global features to improve image reconstruction. In addition, we introduce an entropy-based image conversion method. Extensive evaluations show that the proposed model outperforms benchmark models in anomaly detection and classification using an encoder. The supervised and unsupervised models show improved performances when trained with data preprocessed using the proposed image conversion method.

• Korean Journal of Health Education and Promotion
• /
• v.23 no.3
• /
• pp.17-35
• /
• 2006

Objectives: This study was conducted to investigate the determining factors that have influence on elementary students who go without breakfast and identified their physical, mental and social health that may resulted by skipping breakfast. Methods: For data collection, total 855 elementary students at the 5th and 6th grade in an urban and a rural area were asked to join the survey using a self-administered questionnaire from March 25 to April 9, 2005. The data were analyzed through a series of statistical methods including frequency, percentage, chi-square$(X^2)$ test, t-test and multiple logistic regression analysis. Results: It was found that 24.0 % of respondents would skip breakfast, and influential socioeconomic factors on their skipped breakfast were economic level and mother's occupation. A group that has meals more with family everyday and a group that has breakfast served by mother as usual showed significantly low percentage of going without breakfast at statistical level(p<0.05). In case that other family members used to go without breakfast, there was a tendency to significantly high rate of skipped breakfast(p<0.01). And a group who got up early in the morning and a group who scored high points in the assessment of dietary habit showed significantly low percentage of skipped breakfast(p<0.01). In the result of multiple logistic regression analysis, the significant variables in urban area were economic level, person setting up breakfast, getting-up time and family members who would skip a meal. In rural area, meal-skipping family member, person setting up breakfast and dietary habit were significant variables(p<0.05). In physical influences according to skipping breakfast, students would skip breakfast had more stomach or dizziness(p<0.05). Depression score were higher and school life score were lower in group skipping breakfast, but not statistically significant. Conclusions: In conclusion, in order to reduce the rate of elementary students who often go without breakfast, it is favorable that they should be instructed in the importance of breakfast and the harm of skipped meal, so that they can cultivate positive living habits and dietary attitudes as well. And it is necessary that this health education cover not only students but also all of their family members(especially mother). Furthermore, it is necessary to develop systematic programs for them to get over environment factors.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.12C
• /
• pp.1224-1231
• /
• 2006

This paper proposes a fast encoding algorithm of H.264/AVC multi-view video coding. The amount of data to be encoded for a multi-view video is much more than normal video's data, and the amount of information to be predicted is enormous because of the multi-view video coding uses inter-disparity prediction in addition to inter-motion prediction in conventional video coding. We noticed through an experiment that the efficiency of prediction is getting better in order of intra, inter-disparity, inter-motion, and inter-skip, and proposes a early termination algorithm by means of estimate the adaptive threshold within a GOP unit. In the experiments, the proposed algorithm shows improved processing speed about 32% compared to existing method, and increased amount of bits and distortions are relatively disregardable.

• 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 information and communication convergence engineering
• /
• v.7 no.4
• /
• pp.496-500
• /
• 2009

In EBCOT, the context modeling process takes excessive calculation time and this paper proposed a method to reduce this calculation time. That is, if the finest resolution coefficient is less than a pre-defined transfer factor the coefficient and its descendents skip the context modeling process. There is a trade-off relationship between the calculation time and the image quality or the amount of output data such that as this threshold value increases, the calculation time and the amount of output data decreases, but the image degradation increases. The experimental results showed that in this range the resulting reduction rate in calculation time was from 3% to 64% in average, the reduction rate in output data was from 32% to 73% in average.

• Information and Communications Magazine
• /
• v.25 no.12
• /
• pp.10-18
• /
• 2008

In this paper, we present a low-power design of H.264 codec based on dedicated hardware and software solution on EMP(ETRI Multi-core platform). The dedicated hardware scheme has reducing computation using motion estimation skip and reducing memory access for motion estimation. The design reduces data transfer load to 66% compared to conventional method. The gate count of H.264 encoder and the performance is about 455k and 43Mhz@30fps with D1(720x480) for H.264 encoder. The software solution is with ASIP(Application Specific Instruction Processor) that it is SIMD(Single Instruction Multiple Data), Dual Issue VLIW(Very Long Instruction Word) core, specified register file for SIMD, internal memory and data memory access for memory controller, 6 step pipeline, and 32 bits bus width. Performance and gate count is 400MHz@30fps with CIF(Common Intermediated format) and about 100k per core for H.264 decoder.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2483-2485
• /
• 2003

As design a network system that should command and response to do tasks by PLC, to make factory automation for being distributed facilities which are not guaranteed procedure, through polling method, commands are send orderly by PLC and responses are answered by each sub-unit. It make a guarantee of sequence and drive a master-slave operation in network. This method help to distribute procedure systems and make possible dividing tasks to each unit. To make this method, network protocol's format have Master or Slave ID and tasking data and Master's command or Slave unit's report. In network interpret algorithm, skip algorithm are included and it make guarantee of protecting data though network are in noise. Data that are made by this method are send to PC to transmit long-distance user by internet.

• Journal of the Korean Society for Precision Engineering
• /
• v.20 no.2
• /
• pp.120-128
• /
• 2003

When considering the use of rapid prototyping (RP), there are many issues a designer has to address for handling an STL model, the de facto standard fur RP. Today designers can skip all these issues by visiting web-based service bureaus that readily supply needed information for the RP services. Since orders are taken for RP parts through the web page of service providers designers are now asked to upload their STL files to the company server either by direct upload, ftp file transfer, or as an e-mail attachment. If the service bureau, however, fixes or edits an STL filceto optimize the RP process but neglects to tell its customer about the rework in detail, it may cause problems down the line in processing of the original CAD data for other applications. In this paper, we propose a framework for a collaborative virtual environment between CAD designers and RP processes on the internet which directly provides designers with an advanced preprocessor functionality, design visualization, as well as model display, repair, and slicing over the network. This can help smooth data transfer from CAD to RP process with minimum inconsistency in CAD.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.7 no.8
• /
• pp.66-75
• /
• 1999

In two-stroke engines, methods of evaluating scavenging effciency have been studied in parallel with trials for improvement of scavenging process. But simulating methods have a wide difference with difference with scavenging process in real engines and methods of evaluating scavening efficiency in real engines are very difficult and very expensive. This study shows that there is a possbility of evaluating scavenging efficiency in real engines very easily by analysis of in-cylinder pressure data. And as a characteristics two-stroke engines, the poly tropic indices in the process of compression are varies with degreeof scavenging and good representatives of scavenging efficiency.

• Journal of Broadcast Engineering
• /
• v.11 no.1 s.30
• /
• pp.116-129
• /
• 2006

The progress of data transmission technology through the Internet has spread a variety of realistic contents. One of such contents is multi-view video that is acquired from multiple camera sensors. In general, the multi-view video processing requires encoders and decoders as many as the number of cameras, and thus the processing complexity results in difficulties of practical implementation. To solve for this problem, this paper considers a simple multi-view system utilizing a single encoder and a single decoder. In the encoder side, input multi-view YUV sequences are combined on GOP units by a video mixer. Then, the mixed sequence is compressed by a single H.264/AVC encoder. The decoding is composed of a single decoder and a scheduler controling the decoding process. The goal of the scheduler is to assign approximately identical number of decoded frames to each view sequence by estimating the decoder utilization of a Gap and subsequently applying frame skip algorithms. Furthermore, in the frame skip, efficient frame selection algorithms are studied for H.264/AVC baseline and main profiles based upon a cost function that is related to perceived video quality. Our proposed method has been performed on various multi-view test sequences adopted by MPEG 3DAV. Experimental results show that approximately identical decoder utilization is achieved for each view sequence so that each view sequence is fairly displayed. As well, the performance of the proposed method is examined in terms of bit-rate and PSNR using a rate-distortion curve.