• Journal of Broadcast Engineering
• /
• v.27 no.3
• /
• pp.283-294
• /
• 2022

The feature map used in the network for deep learning generally has larger data than the image and a higher compression rate than the image compression rate is required to transmit the feature map. This paper proposes a method for transmitting a pyramid feature map with high compression rate, which is used in a network with an FPN structure that has robustness to object size in deep learning-based image processing. In order to efficiently compress the pyramid feature map, this paper proposes a structure that predicts a pyramid feature map of a level that is not transmitted with pyramid feature map of some levels that transmitted through the proposed prediction network to efficiently compress the pyramid feature map and restores compression damage through the proposed reconstruction network. Suggested mAP, the performance of object detection for the COCO data set 2017 Train images of the proposed method, showed a performance improvement of 31.25% in BD-rate compared to the result of compressing the feature map through VTM12.0 in the rate-precision graph, and compared to the method of performing compression through PCA and DeepCABAC, the BD-rate improved by 57.79%.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.425-428
• /
• 1999

Turbo decoder were shown to achieve performance within 0.7㏈ of the Shannon capacity limit. This constituted a significant gain in power efficiency over other coding techniques known at the time. In this paper, Turbo code with constraint length K＝4, code rate 1/3, frame size 196bits(6 tail bits), 20㎳ frame and 6bit MAP decoder is implemented using VHDL. The designed Turbo code is used for voice service. Interactions of the system are used to attain large performance improvements.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.627-630
• /
• 1999

For depth extraction from the focus and recovery the shape, determination of criterion function for focus measure and size of the criterion window are very important. However, Texture, illumination, and magnification have an effect on focus measure. For that reason, depth map has a partial high and low peak. In this paper, we propose a depth extraction method from focused images using the error interpolation. This method is modified the error depth into mean value between two normal depth in order to improve the depth map.

• Journal of applied mathematics & informatics
• /
• v.24 no.1_2
• /
• pp.411-419
• /
• 2007

In this paper, $\tilde{2}$-essential rooted maps on the Klein bottle are counted and an explicit expression with the size as a parameter is given. Further, the numbers of singular maps and the maps with one vertex on the Klein bottle are derived.

• 한국정보통신설비학회:학술대회논문집
• /
• 2008.08a
• /
• pp.563-565
• /
• 2008

In this paper, we construct an efficient fuzzy identity-based encryption scheme in the random oracle model. The fuzzy identity-based encryption is an extension of identity-based encryption schemes where a user's public key is represented as his identity. Our construction requires constant number of bilinear map operations for decryption and the size of private key is small compared with the previous fuzzy identity-based encryption of Sahai-Waters. We also presents that our fuzzy identity-based encryption can be converted to attribute-based encryption schemes.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.10a
• /
• pp.663-665
• /
• 2010

By the curing and transformation experiment, it was found thatthe genes of Pseudomonas sp.KU171(pKU19) for MCPA-degrading were located on a plasmid pKU19. Also the plasmid had degradative gens for 2,4-D, 3CB, and DCP. Molecular size of pKU19 was measured to be 31.2Kb. The restriction pattern were analyzed with Eco RI, BglII,XhoI, and the restriction map was generated.

• Science of Emotion and Sensibility
• /
• v.2 no.1
• /
• pp.129-135
• /
• 1999

본 연구는 3D Visual and Auditory Environment Generator(VAEG)에 의해 표현되는 가상환경에서 시각적 현실감의 정도를 평가하고자 하는데 그 목적이 있다. VAEG는 인간의 감성과 일치하는 제품을 설계할 수 있는 prototyping simulator의 일부이다. 계층적인 가상환경을 구성하는 요소로써 size, length, distance 세가지 요소를 선정하여, estimating size/distance, sketching map, 그리고 searching objects와 같은 다양한 방법을 통해 가상환경과 실제환경에서의 시각적인 현실감 차이를 비교하고자 하였다. 실험은 피험자들이 가상환경을 항해하고, 과제를 수행한 후에 설문에 응답하도록 하는 방식으로 진행되었으며, sketched map과 설문서 간의 상관관계를 통해 가상현실시스템의 현실감을 향상시킬 수 있는 보완요소들을 추출할 수 있도록 설계되었다.

• Korean Journal of Remote Sensing
• /
• v.34 no.4
• /
• pp.661-670
• /
• 2018

This study aims to develop a methodology of convolutional neural networks (CNNs) to produce thematic maps from remote sensing data. Optimizing the image size for CNNs was studied, since the size of the image affects to accuracy, working as hyper-parameter. The selected study area is Mt. Ung, located in Dangjin-si, Chungcheongnam-do, South Korea, consisting of both coniferous forest and deciduous forest. Spatial structure analysis and the classification of forest type using CNNs was carried in the study area at a diverse range of scales. As a result of the spatial structure analysis, it was found that the local variance (LV) was high, in the range of 7.65 m to 18.87 m, meaning that the size of objects in the image is likely to be with in this range. As a result of the classification, the image measuring 15.81 m, belonging to the range with highest LV values, had the highest classification accuracy of 85.09%. Also, there was a positive correlation between LV and the accuracy in the range under 15.81 m, which was judged to be the optimal image size. Therefore, the trial and error selection of the optimum image size could be minimized by choosing the result of the spatial structure analysis as the starting point. This study estimated the optimal image size for CNNs using spatial structure analysis and found that this can be used to promote the application of deep-learning in remote sensing.

• Journal of the Korean Institute of Gas
• /
• v.18 no.4
• /
• pp.86-94
• /
• 2014

To develop a Root Cause Analysis Map which determines the cause of the accident in chemical laboratory, The Root Cause Analysis(RCA) Map for the laboratory areas was sketched from Phase 1 of the accident element to Phase 3 of the accident element, based on the RCA Map which is applied in the petrochemical industry. On the basis of laboratory RCA Map which was classified by using such method. The root causes of the 211 accident cases in laboratories were classified from Phase 4 to Phase 5 by the Cause Factor Charting technique and The cause of the accident data were inputted to EXCEL program. After that, The causes of the accident data were sorted and classified by type and each step. So 'Approximate Primary RCA Map Draft' was written. In addition, it was reaffirmed whether the root causes of 211 accidents of laboratory were appropriate to 'Primary RCA Map Draft'. By complementing the cause which was expected to cause future accidents, the RCA Map for chemical laboratories was developed. Based on 'RCA Map' proposed in this study, the causes of accidents were analysed management systems 35%, monitoring 12.2%, Human Factor Eng. 15.1% and education training 12.1% by the size of the frequency from Phase 1 to Phase 5.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.7 no.5
• /
• pp.955-960
• /
• 2012

In this paper, we analysis a competition to organize knowledge map as initial prepare stage based on intellectual property. We also propose the method that can be used to constitute knowledge map related in area to use basic data. In order to extract analysis data, we choose LED as an analysis technology. We provide the result of total analysis for LED application area through analysis of patent level after we extract recent patent ratio, activity index, patent family size, cites per patent as center of patent data.