• Smart Media Journal
• /
• v.4 no.3
• /
• pp.50-55
• /
• 2015

In this paper, we present background modeling method based on Gaussian mixture model to subtract background for night-time video surveillance. In night-time video, it is hard work to distinguish the object from the background because a background pixel is similar to a object pixel. To solve this problem, we change the pixel of input frame to more advantageous value to make the Gaussian mixture model using scaled histogram stretching in preprocessing step. Using scaled pixel value of input frame, we then exploit GMM to find the ideal background pixelwisely. In case that the pixel of next frame is not included in any Gaussian, the matching test in old GMM method ignores the information of stored background by eliminating the Gaussian distribution with low weight. Therefore we consider the stacked data by applying the difference between the old mean and new pixel intensity to new mean instead of removing the Gaussian with low weight. Some experiments demonstrate that the proposed background modeling method shows the superiority of our algorithm effectively.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.6
• /
• pp.228-237
• /
• 2013

In this paper, we propose a normalized cross correlation(NCC)-based multiview background subtraction method which is robust when an object and background have similar color. When the background of the capturing environment is not artificially composed, the regions in the background images which would be occluded by an object tends to have difference colors. The colors of those regions, however, becomes similar when an object enters the capturing environment. Based on this assumption, this paper proposes a concept of GoNCC(Graph of Normalized Cross Correlation). GoNCC is the distribution of NCC between a pixel in an image and pixels related by epipolar constraints with the pixel. The proposed multiview background subtraction method is performed by comparing GoNCC of the current images with the background images. To reduce computational complexity, we perform multiview background subtraction only to the pixels undetermined by single view background subtraction. Experimental results show that the proposed method is more robust to color similarity between an object and background than a single-view background subtraction method and a previous multiview background subtraction method.

• 한국HCI학회:학술대회논문집
• /
• 2009.02a
• /
• pp.631-636
• /
• 2009

Foreground segmentation methods have steadily been researched in the field of computer vision. Especially, background subtraction which extracts a foreground image from the difference between the current frame and a reference image, called as "background image" have been widely used for a variety of real-time applications because of low computation and high-quality. However, if the background scene was dynamically changed, the background subtraction causes lots of errors. In this paper, we propose an efficient background subtraction method in dynamic environment with both static and dynamic scene. The proposed method is a hybrid method that uses the conventional background subtraction for static scene and depth information for dynamic scene. Its validity and efficiency are verified by demonstration in dynamic environment, where a video projector projects various images in the background.

• Journal of Korea Multimedia Society
• /
• v.12 no.10
• /
• pp.1418-1426
• /
• 2009

There has been a lot of interest in an effective method for background subtraction in an effort to separate foreground objects from a predefined background image. Promising results on background subtraction using statistical methods have recently been reported are robust enough to operate in dynamic environments, but generally require very large computational resources and still have difficulty in obtaining clear segmentation of objects. We use a simple running-average method to model a gradually changing background, instead of using a complicated statistical technique. We employ a single global threshold vector, optimized by a genetic algorithm, instead of pixel-by-pixel thresholds. A new fitness function is defined and trained to evaluate segmentation result. The system has been implemented on a PC with a webcam, and experimental results on real images show that the new method outperforms an existing method based on a mixture of Gaussian.

• The KIPS Transactions:PartB
• /
• v.15B no.1
• /
• pp.25-36
• /
• 2008

In this paper, we present a robust hand recognition approach to sudden illumination changes. The proposed approach constructs a background model with respect to hue and hue gradient in HSI color space and extracts a foreground hand region from an input image using the background subtraction method. Eighteen features are defined for a hand pose and multi-class SVM(Support Vector Machine) approach is applied to learn and classify hand poses based on eighteen features. The proposed approach robustly extracts the contour of a hand with variations in illumination by applying the hue gradient into the background subtraction. A hand pose is defined by two Eigen values which are normalized by the size of OBB(Object-Oriented Bounding Box), and sixteen feature values which represent the number of hand contour points included in each subrange of OBB. We compared the RGB-based background subtraction, hue-based background subtraction and the proposed approach with sudden illumination changes and proved the robustness of the proposed approach. In the experiment, we built a hand pose training model from 2,700 sample hand images of six subjects which represent nine numerical numbers from one to nine. Our implementation result shows 92.6% of successful recognition rate for 1,620 hand images with various lighting condition using the training model.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.8 no.6
• /
• pp.957-966
• /
• 2018

Recently, the demand for remote patient monitoring based on IoT has been increased due to aging population and an increase in single-person household. A non-contact biological signal measurement system using multiple IR-UWB radars for remote patient monitoring is proposed in this paper. To reduce error signals, a multilayer Subtraction algorithm is applied because when the background subtraction algorithm was applied to the biological signal processing, errors occurred such as voltage noise and staircase phenomenon. Therefore, a multilayer background subtraction algorithm is applied to reduce error occurrence. The multilayer background subtraction algorithm extracts the signal by calculating the amount of change between the previous clutter and the current clutter. In this study, the SVD algorithm is used. We applied the improved multilayer background subtraction algorithm to biological signal measurement and computed the respiration rate through Fast Fourier Transform (FFT). To verify the proposed system using IR-UWB radars and multilayer background subtraction algorithm, the respiration rate was measured. The validity of this study was verified by obtaining a precision of 97.36% as a result of a control experiment with Neulog's attachment type breathing apparatus. The implemented algorithm improves the inconvenience of the existing contact wearable method.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.11a
• /
• pp.13-16
• /
• 2016

본 논문에서는 순환 신경망을 이용하여 동영상에서의 배경과 전경을 구분하는 알고리즘을 제안한다. 순환 신경망은 일련의 순차적인 입력에 대해서 내부의 루프(loop)를 통해 이전 입력에 의한 정보를 지속할 수 있도록 구성되는 신경망을 말한다. 순환 신경망의 여러 구조들 가운데, 우리는 장기적인 관계에도 반응할 수 있도록 장단기 기억 신경망(Long short-term memory networks, LSTM)을 사용했다. 그리고 동영상에서의 시간적인 연결 뿐 아니라 공간적인 연관성도 배경과 전경을 판단하는 것에 영향을 미치기 때문에, 공간적 순환 신경망을 적용하여 내부 신경망(hidden layer)들의 정보가 공간적으로 전달될 수 있도록 신경망을 구성하였다. 제안하는 알고리즘은 기본적인 배경차분 동영상에 대해 기존 알고리즘들과 비교할만한 결과를 보인다.

• KIISE Transactions on Computing Practices
• /
• v.24 no.3
• /
• pp.151-156
• /
• 2018

In this paper, a method to classify objects in outdoor CCTV images using Convolutional Neural Network(CNN) and background subtraction is proposed. Object candidates are extracted using background subtraction and they are classified with CNN to detect objects in the image. At the end, computation complexity is highly reduced in comparison to other object detection algorithms. A database is constructed by filming alleys and playgrounds, places where crime occurs mainly. In experiments, different image sizes and experimental settings are tested to construct a best classifier detecting person. And the final classification accuracy became 80% for same camera data and 67.5% for a different camera.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.513-518
• /
• 2006

본 논문에서는 특별한 스튜디오 조명이 아닌 가정용 혹은 학교용 조명 하에서 동작하는 감정 표현이 가능한 그림자 놀이 도구를 소개한다. 이 놀이는 사용자의 실사로부터 추출된 가상 그림자에 사용자의 동작에 따라 실시간으로 반응하는 가상 표정 애니메이션 캐릭터를 투영한다. 이 놀이 도구는 행위전시 등의 공연 환경뿐만 아니라 가정, 학교 등의 일상 환경에서도 동작이 가능하도록 설계되었다. 따라서 가장 일반적인 조명의 하나인 가정용 형광등과 백열등 환경을 가정하였고, 배경에 대해서는 어떤 제약도 가정하지 않았다. 사용자의 실사를 구분하는 과정에는 배경 차분화 기법을 사용하였고, 빛의 산란 및 실제 그림자의 영향 등으로 인한 잡음을 최소화하기 위해 몇 가지 잡음 제거 알고리즘을 결합하여 적용하였다. 이 도구에서 가상 그림자의 표정 애니메이션은 사용자의 손동작에 따라 변화하며, 손동작 및 머리 위치의 추적을 위해 새로이 고안한 자질 기반 추적 기법을 적용하였다. 이 도구는 공연장에서의 인터렉티브 미디어 아트 전시 또는 공연, 아동 대상의 미술 교육, 예술치료 등의 분야에 활용될 수 있을 것으로 판단된다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.05a
• /
• pp.407-409
• /
• 2015

In this paper, a hardware implementation of MOD(Moving Object Detection) algorithm is described, which is based GMM(Gaussian Mixture Model) and background subtraction. The EGML(Effective Gaussian Mixture Learning) is used to model and update background. Some approximations of EGML calculations are applied to reduce hardware complexity, and pipelining technique is used to improve operating speed. Gaussian parameters are adjustable according to various environment conditions to achieve better MOD performance. MOD processor is verified by using FPGA-in-the-loop verification, and it can operate with 109 MHz clock frequency on XC5VSX95T FPGA device.