• Journal of KIISE:Software and Applications
• /
• v.31 no.2
• /
• pp.213-223
• /
• 2004

In this paper, we propose a new face detection method for real-time applications. It is based on the template-matching and appearance-based method. At first, we apply Min-max normalization with histogram equalization to the input image according to the variation of intensity. By applying the PCA transform to both the input image and template, PC components are obtained and they are applied to the LDA transform. Then, we estimate the distances between the input image and template, and we select one region which has the smallest distance. SVM is used for final decision whether the candidate face region is a real face or not. Since we detect a face region not the full region but within the $\pm$12 search window, our method shows a good speed and detection rate. Through the experiments with 6 category input videos, our algorithm shows the better performance than the existing methods that use only the PCA transform. and the PCA and LDA transform.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.11
• /
• pp.457-464
• /
• 2019

The weight of pig is one of the main factors in determining the health and growth state of pigs, their shipment, the breeding environment, and the ration of feed, and thus measuring the pig's weight is an important issue in productivity perspective. In order to estimate the pig's weight by using the number of pig's pixels from images, acquired from a Top-view camera, the posture determining and the head removal from images are necessary to measure the accurate number of pixels. In this research, we propose the fast and accurate method to determine the pig's posture by using a fast image processing technique, find the head location by using a fast deep learning technique, and remove pig's head by using light weighted image processing technique. First, we determine the pig's posture by comparing the length from the center of the pig's body to the outline of the pig in the binary image. Then, we train the location of pig's head, body, and hip in images using YOLO(one of the fast deep learning based object detector), and then we obtain the location of pig's head and remove an outside area of head by using head location. Finally, we find the boundary of head and body by using Convex-hull, and we remove pig's head. In the Experiment result, we confirmed that the pig's posture was determined with an accuracy of 0.98 and a processing speed of 250.00fps, and the pig's head was removed with an accuracy of 0.96 and a processing speed of 48.97fps.

• Journal of ILASS-Korea
• /
• v.4 no.1
• /
• pp.13-18
• /
• 1999

It has been known that spray characteristics have an important effect on the mixture formation and directly influence the engine performances and the emissions. Up to now, the measurement of droplet size is well developed such as PDPA and PMAS though the behavior of small droplets during secondary atomization is not clear. Particle image velocimetry(PIV), a planar measuring technique, is a very efficient tool for studying complicated behavior and a fast and reliable method to track numerous droplets during injection. In this study, two-color scanning PIV is designed to obtain quasi-instantaneous two dimensional velocity data by using he-ion laser, rotating mirror and beam splitter. This PIV method which has high temporal and spatial resolution provides the information about the small complex droplet behavior.

• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.3
• /
• pp.9-15
• /
• 2001

Very low bit rate video coding uses the inter-picture video coding method for high compression. The inter-picture video coding is coded based on the information of the previous frames so any packet loss can lead to reduce the image quality on the transmission. In this paper, we proposed the multi-stream generation method for inter-media synchronization of very low bit rate video based on TCP for reliable transmission. The proposed approach performs a reliable transmission via a TCP based protocol. This method incorporates multi-streams in order to enhance the robustness of delivery and can withstand against network jitter. Moreover, the client bandwidths are fully utilized in a highly efficient way.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.10
• /
• pp.2629-2639
• /
• 1997

The video coding for very low bit-rate has recently received considerable attention, but the conventional coding schemes with block based transform suffer from the blocky effect for the constraints of limited bit-rate. In this paper, we present a video coding system based on wavelet transform and multiresolution motion estimation/compensation for very low bit-rate video. The proposed scheme uses the wavelet transform which is flexible to represent non-stationary image signals and adaptable to the human visual characteristics. The wavelet transformed coefficients are coded by various coding modes in accordance with the sum of absolute error after motion estimation/compensation in wavelet decomposed domain. And simple buffer control technique is applied to handle constant image quality. It is shown that the presented scheme has more acceptable image quality without blocky effects than conventional block based transform video coding.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.36-37
• /
• 2014

기존에 영상 내에 원형 검출 방법으로 가장 널리 사용되는 방법은 허프 변환에 기초한다. 허프 변환은 해석적 곡선의 각 점을 원의 중심 좌표와 반지름으로 매핑 시키는 과정을 포함한다. 이러한 과정은 실행시간을 매우 많이 필요로 하고 또한 응용에 따라서 최적인 원 근사화 방법을 찾는데 문제점을 야기하기도 한다. 본 논문에서는 원형 모양인 광 연결 소자 장치로 제한된 응용환경에 대해 원 검출을 빠른 속도로 탐색하는 방법과 최적인 원 근사화 방법을 제안한다. 제안한 방법은 에지 검출과 검출된 에지를 이용한 중심좌표 및 반지름 탐색 그리고 최적화된 원 근사화 방법으로 구성된다. 모의실험을 통하여 제안한 방법은 기존의 오픈라이브러리로 제공되는 OpenCV의 허프 변환에 의한 방법에 비해 원 검출 및 근사화 방법에 있어 성능을 개선할 수 있음을 보인다.

• The KIPS Transactions:PartB
• /
• v.12B no.6 s.102
• /
• pp.661-670
• /
• 2005

According as services that offer multimedia streams are activated on the Internet, users who use these services we increased extremely. As a result of this, the increase of network traffic was occurred. Also it caused the problems such as the discontinuous playback in case of the s#ream playback and the asynchronization between a video and an audio. To solve these problems, it needs a method that guarantee the stable playback of media streams and a method of media transfer that can interact between a service user and media. Existing related researches has achieved the synchronization through various methods, but did not shown the results of satisfaction in the aspect of interaction. In this paper, we inserted the interactive objects in each media file, designed functions that these objects can use each other's information, and solved the interaction and the synchronization between a video and an audio. Also, we solved the discontinuous playback of a stream by the insufficiency of network bandwidth through using variable buffers.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.40 no.9
• /
• pp.791-799
• /
• 2016

To ensure the reliability and safety of various mechanical systems in accordance with their high-speed usage, it is necessary to evaluate the dynamic deformation behavior of structural materials under impact load. However, it is not easy to understand the dynamic deformation behavior of the structural materials using experimental methods in the high strain-rate range exceeding $10^4\;s^{-1}$. In this study, the Taylor bar impact test was conducted to investigate the dynamic deformation behavior of metallic materials in the high strain-rate region, using a high-speed photography system. Numerical analysis of the Taylor bar impact test was performed using AUTODYN S/W. The results of the analysis were compared with the experimental results, and the material behavior in the high strain-rate region was discussed.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.71-72
• /
• 2014

최근 들어 드라마나 영화 속 인물 들의 얼굴을 다른 얼굴로 교체하고자 하는 시도들이 주목을 받고 있고, 사용자들의 관심 또한 증가하고 있다. 이러한 연구는 오래 전부터 많이 연구되어 왔으나, 조명 변화에 강인한 피부 톤 변환 문제는 여전히 주요 이슈로 자리하고 있다. 본 논문에서는 히스토그램 명세화 기법을 응용하여 피부 톤을 자연스럽게 변환 시켜줄 수 있는 기법을 제안한다. 두 얼굴 영상의 R, G, B 채널에 대한 히스토그램을 각각 생성하고 이들을 누적분포함수로 각각 변환한다. 변환된 누적분포함수들을 이용하여 두 얼굴 영상의 히스토그램을 근사 시킬 수 있는 변환 표를 생성하고 이를 이용하여 얼굴 영상의 피부 톤을 변환시킨다. 조명 변화로 인해 발생하는 전반사 영역은 선형 보간 법을 사용하여 제거한다. 실험 결과, 제안하는 방법이 조명 변화에 적응적으로 대처하면서 자연스럽게 피부 톤을 변환 시켜줄 수 있음을 확인하였다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2002.05d
• /
• pp.781-785
• /
• 2002

다양한 정보들을 디지털화하여 여러 형태의 미디어로 표현함으로 인해 정보의 저장과 관리가 어렵고 이로 인해 자료의 처리효율이 떨어지거나 상호 교환과 편집 등이 용이하지 않다. 본 논문에서는 동영상 강의를 효율적으로 처리하기 위해 주 제어기, 어휘분석기, 마우스 및 구문 분석기, 변수 테이블, 명령어 관리기, 수치함수 관리기, 변수 관리기, 스크립터 편집기, 스크립터 브라우저 등을 체계적으로 정립하여 다양한 미디어 매체를 하나의 파일형식으로 변환함으로서 동영상 강의 시스템에서 미디어 자료 관리를 원활하게 하고, 미디어의 처리속도와 시스템의 부하를 줄이는 미디어 통합모듈의 알고리즘을 설계하고 구현한다.