• 전자공학회논문지CI
• /
• 제40권3호
• /
• pp.137-143
• /
• 2003

Key frame extraction has been recognized as one of tile important research issues in video information retrieval. Although progress has been made in key frame extraction, the existing approaches do not evaluate the importance of extracted frame Quantitatively, and Dey are computationally expensive or ineffective. In this Paper, we introduce a new 미해rithm for key frame extraction using shot coverage and distortion. The algorithm finds significant key frames from candidate key frames. When selecting the candidate frames, the coverage rate to the total frames from each frame is computed by using the difference between the adjacent frames in tile shot. The frames within 10$\%$ from the top are selected as the candidates. Then, by comfuting the distortion rate of the candidates against all frames, pick the frames which arc most representative are selected. The performance of the proposed algorithm has been verified by a statistical test. Experiments show that more than 13-50$\%$ improvement has been obtained by the proposed algorithm compared to the existing methods.

• 한국컴퓨터정보학회논문지
• /
• 제20권12호
• /
• pp.21-28
• /
• 2015

In this paper, we propose a method of glasses detection in facial image. we develop a detection method of glasses with a weighted sum of the results that detected by facial element detection and glasses frame candidate region. Component of the face detection method detects the glasses, by defining the detection probability of the glasses according to the detection of a face component. Method using the candidate region of the glasses frame detects the glasses, by defining feature of the glasses frame in the candidate region. finally, The results of the combined weight of both methods are obtained. The proposed method in this paper is expected to increase security system's recognition on facial accessories by raising detection performance of glasses or sunglasses for using ATM.

• 한국위성정보통신학회논문지
• /
• 제9권4호
• /
• pp.1-6
• /
• 2014

컬러 영상 처리 기반의 연기 검출은 화재의 조기 검출에 적합한 검출 대상이다. 연기 검출을 위한 방법으로 움직임과 색상이 전처리로서 처리되며, 확산, 질감, 형태, 방향성 등의 성질이 후처리로서 사용된다. 본 논문은 연기의 특성 중 밀도적인 분포 특성 검출 방법을 제안한다. 연기의 움직임을 10Frame 간격으로 1초 동안 축적한 이미지에 색상을 문턱치 처리해 후보영역을 생성하고, OBTP(Object Block Ternary Pattern)을 적용해 연기의 패턴임을 확인한다. 모든 처리는 Block 기반으로, 움직임 검출은 차분 영상에 적응 문턱치를 적용해 움직이는 물체의 후보영역을 결정했다. 결정된 후보영역을 1초간 축적하고 연기 색상의 문턱치 조건을 적용한다. 각각의 연기 후보 영역을 특정 위치의 16개 Block 값을 중앙 Block 값과 비교하고 삼진화 된 패턴을 연기의 패턴과 비교하여 연기를 결정한다.

• International journal of advanced smart convergence
• /
• 제8권2호
• /
• pp.116-125
• /
• 2019

The fire should extinguish as soon as possible because it causes economic loss and loses precious life. In this study, we propose a new atypical fire and smoke detection algorithm using deep learning and color histogram of fire and smoke. First, input frame images obtain from the ONVIF surveillance camera mounted in factory search motion candidate frame by motion detection algorithm and mean square error (MSE). Second deep learning (Faster R-CNN) is used to extract the fire and smoke candidate area of motion frame. Third, we apply a novel algorithm to detect the fire and smoke using color histogram algorithm with local area motion, similarity, and MSE. In this study, we developed a novel fire and smoke detection algorithm applied the local motion and color histogram method. Experimental results show that the surveillance camera with the proposed algorithm showed good fire and smoke detection results with very few false positives.

• 전자공학회논문지B
• /
• 제30B권12호
• /
• pp.21-30
• /
• 1993

In this paper, we propose a connected Korean digit recohnition system employing neural networks and lexical constraints of the Korean digits. In the proposed recognition system, firstly, each frame of digit string is labelled by phoneme classification neural networks.which are trained with the reference phoneme segments extracted form an isolated digit based on the position information. And, the frame labels are combined with each other for constructing the phoneme segments. Then, these segments are combined to form a digit candidate using the digit combination rules. The digit candidate is decided based on the condition for digit decision. If the condition is not satisfied, the digit candidate is further recognized using the digit decision neural network in the next step. In our approach, the neural networks are trained with 10 isolated digits uttered by 5 male speakers. To investigate the performance of the proposed recognition system, an intensive computer simulation on the 30 connected digit strings uttered by 5 male speakers is performed. The simulation result indicates that 95.6% digit recognition rate and 82% digit string recognition rate are provided by the proposed Korean digit recognition system.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권3호
• /
• pp.1377-1389
• /
• 2016

This paper presents a target model update scheme for the mean-shift tracking with background weighted histogram. In the scheme, the target candidate histogram is corrected by considering the back-projection weight of each pixel in the kernel after the best target candidate in the current frame image is chosen. In each frame, the target model is updated by the weighted average of the current target model and the corrected target candidate. We compared our target model update scheme with the previous ones by applying several test sequences. The experimental results showed that the object tracking accuracy was greatly improved by using the proposed scheme.

• 한국복합재료학회:학술대회논문집
• /
• 한국복합재료학회 2004년도 추계학술발표대회 논문집
• /
• pp.232-235
• /
• 2004

This is the preliminary study to develop composite window frame of commercial aircraft using VaRTM process. For two candidate carbon fabric(triaxial fabric, sleeving braider dry carbon fiber), specimens were fabricated using VaRTM process ,and then the physical & mechanical tests were performed to gain material property according to ASTM. FEM analysis for each candidate carbon fabric were performed to find the minimum ply number and weight for composite window frame. In this study Tsai-Wu strength failure criteria was utilized to evaluate the safety of structure.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 추계종합학술대회 논문집(3)
• /
• pp.137-140
• /
• 2000

In this paper, we propose as new algorithm for the selection of key frames in a given video. For the selected key frames to be well defined, the selected key frames need to spread out on the whole temporal domain of the given video and guaranteed not to be duplicate. For this purpose, we take the first frame of each shot of the video as the candidate key frame to represent the video. To reduce the overall processing time, we eliminate some candidate key frames which are visually indistinct in the histogram difference. The key frames are then selected using a clustering processing based on the singly linked hierarchical tree. To make the selected key frames be distributed evenly on the whole video, the deviation and time difference between the selected key frames are used. The simulation results demonstrate that our method provides the better performance compared with previous methods.

• Composites Research
• /
• 제19권6호
• /
• pp.1-7
• /
• 2006

본 논문은 수지충전공정을 활용한 상업용 항공기의 복합재 윈도우 프레임을 개발하기 위한 선행 연구이다. 윈도우 프레임을 제작하기 위해 고려중인 2종류의 카본 섬유-triaxial overbraid 와 sleeving braider-에 대해서 VaRTM 공정을 활용한 시편을 제작하고, ASTM 시험법에 따라 물성치를 얻기 위해 물리적/기계적 성질 시험이 실시되었다. 이후 복합재 윈도우 프레임의 최소 플라이수와 무게를 얻기 위해 각각의 탄소 섬유에 대하여 유한요소해석이 수행되었다. 본 연구에서는 구조물의 안전성을 평가하기 위하여 Tsai-Wu 파손강도이론을 사용했다.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제10권3호
• /
• pp.273-281
• /
• 2004

동영상의 현재 블록의 움직임 벡터와 이전 블록의 움직임 벡터는 시간적 상관성을 갖고 있다. 본 논문에서는 영상의 시간적인 특성과 움직임 벡터의 가운데 중심 분포 특성을 이용하는 예측 탐색 알고리즘을 제안한다. 제안된 알고리즘은 이전 프레임 블록으로부터 예측된 움직임 벡터와 분할된 탐색 구간에 속하는 후보 벡터 중에서 가장 작은 SAD 값을 갖는 점을 정확한 움직임 벡터를 찾기 위한 초기 탐색점 위치로 결정한다. 실험 결과 제안된 방식은 FS를 제외한 기존의 대표적인 고속 탐색 방식들에 비해 PSNR 값에 있어서 평균적으로 0.19∼0.46㏈ 개선되고 영상에 따라 최고 1.06㏈ 정도 우수한 결과를 나타내었다.