• 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.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2007년도 춘계종합학술대회
• /
• pp.145-148
• /
• 2007

논문은 비디오 픽쳐 타입을 이용한 비디오 셧 검출에 관한 연구이다. 제안한 방법은 압축된 비디오 프레임에 대하여 원 영상을 복원하지 않고, 압축 상태의 비디오 프레임을 이용한다. I픽쳐 프레임에서 DC영상을 복원하고, P픽쳐 프레임에서는 매크로블록의 개수를 이용하여 비디오 셧을 검출 한다. 테스트 비디오를 이용하여 실험 결과 $85\sim98%$의 장면전환 검출이 가능 하였고, 압축비트스트림을 복원하여 장면전환의 셧을 검출 하는 기법에 비해 4배 빠른 검색이 가능하다.

• 제어로봇시스템학회논문지
• /
• 제8권3호
• /
• pp.275-280
• /
• 2002

A new algorithm for gradual scene change detection in MPEG based frame sequences is proposed in this paper. The proposed algorithm is based on the fact that most of gradual curves can be characterized by variance distributions of edge information in the frame sequences. Average edge frame sequences are obtained by performing "sober" edge detection. Features are extracted by comparing variances with those of local blocks in the average edge frames. Those features are further processed by the opening operation to obtain smoothing variance curves. The lowest variance in the local frame sequences is chosen as a gradual detection point. Experimental results show that the proposed method provides 85% precision and 86% recall rate fur gradual scene changes.

• 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.

• 방송공학회논문지
• /
• 제27권6호
• /
• pp.897-905
• /
• 2022

비디오 영상 내 주요 프레임(Key Frame) 검출은 컴퓨터 비전 분야에서 꾸준히 연구되고 있는 분야 중 하나이다. 최근 심층학습(Deep Learning) 기술의 발전으로 비디오 영상에서의 주요 프레임 검출 성능이 향상 되었으나, 다양한 종류의 영상 콘텐츠 및 복잡한 배경으로 인해 여전히 효과적인 학습이 어려운 문제점이 있다. 본 논문에서는 대조적 학습(Contrastive Learning)과 메모리 뱅크(Memory Bank)를 통해 영상의 주요 프레임을 검출하는 새로운 방법을 제안한다. 제안하는 방법은 입력 프레임과 같은 영상 내 이웃하는 프레임 간 차이와 다른 영상 내 프레임과의 차이를 기반으로 특징 추출 신경망을 학습한다. 이와 같은 대조적 학습을 통해 메모리 뱅크에 주요 프레임을 저장 및 갱신하여 영상의 중복성을 효과적으로 제거한다. 비디오 영상 데이터셋에서의 실험 결과를 통해 제안하는 방법의 성능을 검증하였다.

• Journal of information and communication convergence engineering
• /
• 제5권3호
• /
• pp.269-272
• /
• 2007

The shortcut shot detection based on MPEG compressed video bitstream is presented in this paper. The detection algorithm is used the video picture frame from MPEG compressed video directly not to be decompressed the original image. For shortcut detection, I and P frame of MPEG video bitstream are classified. The changing scene cuts at I pictures are detected by the decompressed DC image and scene cuts at P picture frame by monitoring the percentage of intra-macroblocks per P picture are detected. Experimental results using test video bitstream QVGA results in average 92% detection rate, searching time is taken around 4.5 times faster in comparison with changing scene shot detection algorithm which is decompressed the compressed bitstream.

• Smart Structures and Systems
• /
• 제28권6호
• /
• pp.811-825
• /
• 2021

High-voltage isolating switches play a paramount role in ensuring the safety of power supply systems. However, their exposure to outdoor environmental conditions may cause serious physical defects, which may result in great risk to power supply systems and society. Image processing-based methods have been used for anomaly detection. However, their accuracy is affected by numerous uncertainties due to manually extracted features, which makes the anomaly detection of isolating switches still challenging. In this paper, a vision-based anomaly detection method for isolating switches, which uses the rotational angle of the switch system for more accurate and direct anomaly detection with the help of deep learning (DL) and image processing methods (Single Shot Multibox Detector (SSD), improved frame differencing method, and Hough transform), is proposed. The SSD is a deep learning method for object classification and localization. In addition, an improved frame differencing method is introduced for better feature extraction and a hough transform method is adopted for rotational angle calculation. A number of experiments are conducted for anomaly detection of single and multiple switches using video frames. The results of the experiments demonstrate that the SSD outperforms the You-Only-Look-Once network. The effectiveness and robustness of the proposed method have been proven under various conditions, such as different illumination and camera locations using 96 videos from the experiments.

• 한국컴퓨터정보학회논문지
• /
• 제20권7호
• /
• pp.9-16
• /
• 2015

This paper proposes the parallel design of a shot change detection algorithm using frame segmentation and moving blocks. In the proposed approach, the high parallel processing components, such as frame histogram calculation, block histogram calculation, Otsu threshold setting function, frame moving operation, and block histogram comparison, are designed in parallel for NVIDIA GPU. In order to minimize memory access delay time and guarantee fast computation, the output of a GPU kernel becomes the input data of another kernel in a pipeline way using the shared memory of GPU. In addition, the optimal sizes of CUDA processing blocks and threads are estimated through the prior experiments. In the experimental test of the proposed shot change detection algorithm, the detection rate of the GPU based parallel algorithm is the same as that of the CPU based algorithm, but the average of processing time speeds up about 6~8 times.

• 한국컴퓨터정보학회논문지
• /
• 제22권3호
• /
• pp.45-51
• /
• 2017

In this paper, we propose a consecutive-frame super-resolution method to tackle a moving object problem. The super-resolution is a method restoring a high resolution image from a low resolution image. The super-resolution is classified into two types, briefly, single-frame super-resolution and consecutive-frame super-resolution. Typically, the consecutive-frame super-resolution recovers a better than the single-frame super-resolution, because it use more information from consecutive frames. However, the consecutive-frame super-resolution failed to recover the moving object. Therefore, we proposed an improved method via moving object detection. Experimental results showed that the proposed method restored both the moving object and the background properly.

• 제어로봇시스템학회논문지
• /
• 제7권4호
• /
• pp.304-308
• /
• 2001

The increase of video data makes the demand of efficient retrieval, storing, and browsing technologies necessary. In this paper, a video segmentation method (scene change detection method, or shot boundary detection method) for the development of such systems is proposed. For abrupt cut detection, inter-frame similarities are computed using luminance and edge histograms and a cut is declared when the similarities are under th predetermined threshold values. A gradual scene change detection is based on the similarities between the current frame and the previous shot boundary frame. A correlation method is used to obtain universal threshold values, which are applied to various video data. Experimental results show that propose method provides 90% precision and 98% recall rates for abrupt cut, and 59% precision and 79% recall rates for gradual change.