• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.55-58
• /
• 2024

최근 각 기업의 AI 면접시스템 도입이 증가하고 있으며, AI 면접에 대한 실효성 논란 또한 많은 상황이다. 본 논문에서는 AI 면접 과정에서 지원자를 평가하는 방식을 시각, 음성, 자연어처리 3영역에서 구현함으로써, 면접 지원자를 다방면으로 분석 방법론의 적절성에 대해 평가하고자 한다. 첫째, 시각적 측면에서, 면접 지원자의 감정을 인식하기 위해, 합성곱 신경망(CNN) 기법을 활용해, 지원자 얼굴에서 6가지 감정을 인식했으며, 지원자가 카메라를 응시하고 있는지를 시계열로 도출하였다. 이를 통해 지원자가 면접에 임하는 태도와 특히 얼굴에서 드러나는 감정을 분석하는 데 주력했다. 둘째, 시각적 효과만으로 면접자의 태도를 파악하는 데 한계가 있기 때문에, 지원자 음성을 주파수로 환산해 특성을 추출하고, Bidirectional LSTM을 활용해 훈련해 지원자 음성에 따른 6가지 감정을 추출했다. 셋째, 지원자의 발언 내용과 관련해 맥락적 의미를 파악해 지원자의 상태를 파악하기 위해, 음성을 STT(Speech-to-Text) 기법을 이용하여 텍스트로 변환하고, 사용 단어의 빈도를 분석하여 지원자의 언어 습관을 파악했다. 이와 함께, 지원자의 발언 내용에 대한 감정 분석을 위해 KoBERT 모델을 적용했으며, 지원자의 성격, 태도, 직무에 대한 이해도를 파악하기 위해 객관적인 평가지표를 제작하여 적용했다. 논문의 분석 결과 AI 면접의 다면적 평가시스템의 적절성과 관련해, 시각화 부분에서는 상당 부분 정확도가 객관적으로 입증되었다고 판단된다. 음성에서 감정분석 분야는 면접자가 제한된 시간에 모든 유형의 감정을 드러내지 않고, 또 유사한 톤의 말이 진행되다 보니 특정 감정을 나타내는 주파수가 다소 집중되는 현상이 나타났다. 마지막으로 자연어처리 영역은 면접자의 발언에서 나오는 말투, 특정 단어의 빈도수를 넘어, 전체적인 맥락과 느낌을 이해할 수 있는 자연어처리 분석모델의 필요성이 더욱 커졌음을 판단했다.

• Journal of Digital Contents Society
• /
• v.18 no.7
• /
• pp.1323-1331
• /
• 2017

Despite the fact that the copyright has grown into a large-scale business, there are many constant problems especially in image copyright. In this study, we propose an automatic object extraction and classification system for the scanned educational book image by combining document image processing and intelligent information technology like deep learning. First, the proposed technology removes noise component and then performs a visual attention assessment-based region separation. Then we carry out grouping operation based on extracted block areas and categorize each block as a picture or a character area. Finally, the caption area is extracted by searching around the classified picture area. As a result of the performance evaluation, it can be seen an average accuracy of 83% in the extraction of the image and caption area. For only image region detection, up-to 97% of accuracy is verified.

• Smart Media Journal
• /
• v.1 no.2
• /
• pp.12-22
• /
• 2012

Today we are witnessing an increasingly widespread use of cameras in our lives for video surveillance, robot vision, and mobile phones. This has led to a renewed interest in computer vision in general and an on-going boom in human activity recognition in particular. Although not particularly fancy per se, human gait is inarguably the most common and frequent action. Early on this decade there has been a passing interest in human gait recognition, but it soon declined before we came up with a systematic analysis and understanding of walking motion. This paper presents a set of DBN-based models for the analysis of human gait in sequence of increasing complexity and modeling power. The discussion centers around HMM-based statistical methods capable of modeling the variability and incompleteness of input video signals. Finally a novel idea of extending the discrete state Markov chain with a continuous density function is proposed in order to better characterize the gait direction. The proposed modeling framework allows us to recognize pedestrian up to 91.67% and to elegantly decode out two independent gait components of direction and posture through a sequence of experiments.

• Journal of Biosystems Engineering
• /
• v.21 no.3
• /
• pp.325-335
• /
• 1996

Visual features are crucial for monitoring the growth state, indexing the drying performance, and grading the quality of oak mushrooms. A computer vision system with neural net information processing technique was utilized to quantize quality factors of a dried oak mushrooms distributed over the cap and gill sides. In this paper, visual feature extraction algorithm were integrated with the neural net processing to deal with various fuzzy patterns of mushroom shapes and to compensate the fault sensitiveness of the crisp criteria and heuristic rules derived from the image processing results. The proposed algorithm improved the segmentation of the skin features of each side, the identification of cap and gill surfaces, the identification of stipe states and removal of the stipe, etc. And the visual characteristics of dried oak mushrooms were analyzed and primary visual features essential to tile quality evaluation were extracted and quantized. In this study, black and white gray images were captured and used for the algorithm development.

• Cartoon and Animation Studies
• /
• s.51
• /
• pp.293-319
• /
• 2018

Thanks to the advancement of the related technologies and equipment, today's video contents like movies, animations and soap operas are rapidly expanding their expressible cinematic imagination area. In order to fulfill the elevated visual expectations of audiences and realize exciting storytelling and fantastic world, the fusion of different techniques is actively used, and the reality for visual effects and image synthesis is increasing more and more. Accordingly, recent VFX-oriented movies using CG have a much more complicated production process than before. Therefore, the importance of Pre-visualization, aka Pre-vis is becoming bigger in the planning process for sophisticated design. Pre-vis means that the advance visualization for stories or directing ideas in the planning process before starting production of movies or animations. 3D animated Pre-visualization realizing directors' abstract and ambiguous ideas in 3 dimensional environment in advance is, as a powerful means for visual storytelling, briskly used focusing on the VFX film industry on which the present CG is broadly used, and the role of Pre-vis throughout productions has increased compared to the past. The studies, however, on the role and utility of Pre-vis are not enough. Therefore, this study was conducted on the role of Pre-vis used for present VFX movie productions using the examples of 3D animated Pre-visualization production in which the researcher of this study participated. In this study, the role of the Pre-vis that is subdivided presently, is divided into and 3D animatics and their each role is analyzed with the example images. Through this, the characteristics that Pre-vis should have are clarified and the concept of the advantages and utility led by the use of Pre-vis in productions is strengthened. The goal of this study is to induce active uses of Pre-vis throughout productions after forming consensus about the various roles of Pre-vis and their utility.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.30 no.4
• /
• pp.317-323
• /
• 2024

According to the section A-VI/3 of the International Convention on Standards of Training, Certification and Watchkeeping for Seafarers (STCW), Water-based firefighting training center is mandatory to obtain onboard certificates. This space, being similar to fire situations on ships requires that safety measures be quantified to ensure occupant safety and establish operational standards. For fire safety evaluation, cases were designed based on the presence or absence of smoke control equipment using Pyrosim based on Fire Dynamics Simulation (FDS). Vector analysis was performed to evaluate flow of smoke and heat. Available safe escape time / required safe escpae time (ASET/RSET) analysis was conducted to evaluate safety by comparing the interpreted numerical results through Pathfinder. During safety evaluation of the current operational condition, the appropriateness of the function of each smoke control equipment was numerically and visually indicated. The emergency situation with dust collector stopped was expressed by each evacuation time and safety margin of 111.2 seconds, suggesting that be used as a standard of evacuation time.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06b
• /
• pp.400-404
• /
• 2008

무인항공기를 이용한 실시간 공중 모니터링은 재난 재해, 테러 등의 위기상황을 사전에 대비하고, 사고 발생 시 피해상황을 신속하게 파악할 수 있는 효율적인 관리 시스템이다. 실시간 공중 모니터링을 위해 무인항공부문에서는 고성능의 카메라, 관성항법장치, 레이저 스캐너, GPS 수신기 등의 다중 센서들을 장착하고, 제어하며 각 센서들로부터 입력받은 데이터 처리 및 지상국으로 데이터 전송이 실시간으로 가능해야 한다. 기존 무인 모니터링 시스템들은 카메라와 같이 단일 센서의 운용을 목적으로 설계되었으나, 본 연구에서는 레이져 스캐너, 적외선카메라를 포함하는 다중센서를 위한 컴퓨터를 설계하였다. 최근 다중센서를 장착한 관측시스템에 관한 연구가 미국 및 유럽에서 수행되고 있으나, 아직 개발이 완료되지 않은 상태이다. 본 논문에서는 고성능 다중 센서 데이터 처리를 위해 실시간 소프트웨어, 고속 대용량 데이터처리 기술, 고속 압축 기술, 이기종 다중 센서들 간의 시각 동기화 기능을 제공하는 탑재컴퓨터의 설계결과를 소개하였다.

• The Journal of the Korea Contents Association
• /
• v.7 no.6
• /
• pp.169-176
• /
• 2007

Hospital managements are faced with severe competition and need more complicated information systems for its efficiency than before. Hospital managers have to deal with these environments and provide patients with better services than ever to survive. This thesis analyzes and redesigns the hospital administration processes in terms of patient-oriented view. For that purpose we define the processes of the hospital administration using process design tool like ARIS-Toolset. We also introduce basic concepts of Hospital Information Systems(HIS). We finally reviewed the result of reengineering of hospital management process with AHP methodology.

• Communications of Mathematical Education
• /
• v.24 no.3
• /
• pp.709-730
• /
• 2010

The inverse function of a one-to-one correspondence is explained with a graph, a numerical formula or other useful expressions. The purpose of this paper is to know how low achieving students understand the learning contents needed reversible thinking about irrational functions. Low achieving students in this study took paper-pencil test and their written answers were collected. They made various mistakes in solving problems. Their error types were grouped into several classes and identified in this analysis. Most students did not connected concepts that they learned in the lower achieving students to think in reverse order in case of and to visualize concepts of functions. This paper implies that it is very important to take into account students' accommodation and reversible thinking activity.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.1129-1132
• /
• 2019

최근 현대인들은 식습관이 불규칙하고 서구화되면서, 건강상의 많은 문제를 겪고 있다. 이와 더불어 1인 가구의 증가와 간단한 구매 방법 등으로 인해 온라인 몰 사용자가 늘어나고 있다. 본 프로젝트는 이러한 추세를 바탕으로, 사용자가 자주 사용하는 온라인 몰에 축적된 데이터를 기반으로 사용자의 식습관을 분석한다. 뿐만 아니라, 이를 바탕으로 구매 패턴을 분석하여 사용자의 영양 상태를 개선시킬 수 있는 상품 추천 서비스를 제공한다. 사용자는 자주 사용하는 온라인 쇼핑몰에서 상품 구매를 함과 동시에 구매한 상품에 대해 시각화된 영양소 분석 결과와 구매 패턴 분석 결과를 제공받을 수 있다. 본 논문에서는 개발한 API를 통해 사용자는 부족한 영양소를 쉽게 파악하여 효율적으로 건강관리를 할 수 있게 된다. 더 나아가, 자신의 구매 패턴을 파악할 수 있게 되어 현명한 소비 습관을 만드는 데에 기여할 수 있다.