• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.07a
• /
• pp.42-45
• /
• 2010

This paper proposes an intelligent video surveillance system for human object tracking. The proposed system integrates the object extraction, human object recognition, face detection, and camera control. First, the object in the video signals is extracted using the background subtraction. Then, the object region is examined whether it is human or not. For this recognition, the region-based shape descriptor, angular radial transform (ART) in MPEG-7, is used to learn and train the shapes of human bodies. When it is decided that the object is human or something to be investigated, the face region is detected. Finally, the face or object region is tracked in the video, and the pan/tilt/zoom (PTZ) controllable camera tracks the moving object with the motion information of the object. This paper performs the simulation with the real CCTV cameras and their communication protocol. According to the experiments, the proposed system is able to track the moving object(human) automatically not only in the image domain but also in the real 3-D space. The proposed system reduces the human supervisors and improves the surveillance efficiency with the computer vision techniques.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.07a
• /
• pp.40-41
• /
• 2010

실시간 파노라마 영상 합성 기술을 모바일 콘텐츠에 적용하기 위해서는 모바일 디바이스에 적합한 영상 합성 방식을 정의해야 한다. 특징점을 추출하여 연속적인 영상들 사이의 유사성을 찾아내는 방식 [1]은 소모되는 연산비용을 고려할 때 모바일 디바이스에 응용하기 부적합하다. 따라서 기존의 접근방식에 비해 소요되는 연산비용을 줄이고, 성능은 유지하는 기법이 필요하다. 본 논문에서는 특징점의 표현자(descriptor)를 생성하지 않고 특징점(feature point) 주변의 움직임을 추정(motion estimation)하여 연속적인 영상의 위치를 찾아내는 기법을 제안한다.

• Journal of Korea Multimedia Society
• /
• v.17 no.1
• /
• pp.15-25
• /
• 2014

In this paper, we propose a loitering detection using trajectory probability distribution and local direction descriptor for intelligent surveillance system. We use a background modeling method for detecting moving object and extract the motion features from each moving object for making feature vectors. After that, we detect the loitering behavior person using K-Nearest Neighbor classifier. We test the proposed method in real world environment and it can achieve real time and robust detection results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.2
• /
• pp.483-503
• /
• 2014

In this paper, human action recognition using pyramid histograms of oriented gradients and collaborative multi-task learning is proposed. First, we accumulate global activities and construct motion history image (MHI) for both RGB and depth channels respectively to encode the dynamics of one action in different modalities, and then different action descriptors are extracted from depth and RGB MHI to represent global textual and structural characteristics of these actions. Specially, average value in hierarchical block, GIST and pyramid histograms of oriented gradients descriptors are employed to represent human motion. To demonstrate the superiority of the proposed method, we evaluate them by KNN, SVM with linear and RBF kernels, SRC and CRC models on DHA dataset, the well-known dataset for human action recognition. Large scale experimental results show our descriptors are robust, stable and efficient, and outperform the state-of-the-art methods. In addition, we investigate the performance of our descriptors further by combining these descriptors on DHA dataset, and observe that the performances of combined descriptors are much better than just using only sole descriptor. With multimodal features, we also propose a collaborative multi-task learning method for model learning and inference based on transfer learning theory. The main contributions lie in four aspects: 1) the proposed encoding the scheme can filter the stationary part of human body and reduce noise interference; 2) different kind of features and models are assessed, and the neighbor gradients information and pyramid layers are very helpful for representing these actions; 3) The proposed model can fuse the features from different modalities regardless of the sensor types, the ranges of the value, and the dimensions of different features; 4) The latent common knowledge among different modalities can be discovered by transfer learning to boost the performance.

• Journal of Electrical Engineering and Technology
• /
• v.9 no.2
• /
• pp.739-748
• /
• 2014

In this paper, we propose pyramid appearance and global structure action descriptors on both RGB and depth motion history images and a model-free method for human action recognition. In proposed algorithm, we firstly construct motion history image for both RGB and depth channels, at the same time, depth information is employed to filter RGB information, after that, different action descriptors are extracted from depth and RGB MHIs to represent these actions, and then multimodality information collaborative representation and recognition model, in which multi-modality information are put into object function naturally, and information fusion and action recognition also be done together, is proposed to classify human actions. To demonstrate the superiority of the proposed method, we evaluate it on MSR Action3D and DHA datasets, the well-known dataset for human action recognition. Large scale experiment shows our descriptors are robust, stable and efficient, when comparing with the-state-of-the-art algorithms, the performances of our descriptors are better than that of them, further, the performance of combined descriptors is much better than just using sole descriptor. What is more, our proposed model outperforms the state-of-the-art methods on both MSR Action3D and DHA datasets.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.6
• /
• pp.3086-3103
• /
• 2017

Video images captured by vehicle cameras often contain blurry or dithering frames due to inadvertent motion from bumps in the road or by insufficient illumination during the morning or evening, which greatly reduces the perception of objects expression and recognition from the records. Therefore, a real-time electronic stabilization method to correct fuzzy video from driving recorders has been proposed. In the first stage of feature detection, a coarse-to-fine inspection policy and a scale nonlinear diffusion filter are proposed to provide more accurate keypoints. Second, a new antiblurry binary descriptor and a feature point selection strategy for unintentional estimation are proposed, which brought more discriminative power. In addition, a new evaluation criterion for affine region detectors is presented based on the percentage interval of repeatability. The experiments show that the proposed method exhibits improvement in detecting blurry corner points. Moreover, it improves the performance of the algorithm and guarantees high processing speed at the same time.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.16 no.2
• /
• pp.157-175
• /
• 2005

The main purpose of this paper is to explain and analyze visual descriptors of MPEG-7 for representing multimedia content. This study describes MPEG-7 visual descriptors that are made of color, shape, texture, and motion using some examples and application areas in detail. As a result, those visual descriptors can represent the rich and deep features in multimedia contents domain. Also, the use of those descriptors will increase the retrieval effectiveness as well as the interoperability among systems through the consistency of the content representation.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39C no.8
• /
• pp.731-737
• /
• 2014

This paper presents abnormal behavior detection in crowd within surveillance video. We have defined below two cases as a abnormal behavior; first as a sporadically spread phenomenon and second as a sudden running in same direction. In order to detect these two abnormal behaviors, we first extract the motion vector and propose a new descriptor which is combined MHOF(Multi-scale Histogram of Optical Flow) and DCHOF(Directional Change Histogram of Optical Flow). Also, binary classifier SVM(Support Vector Machine) is used for detection. The accuracy of the proposed algorithm is evaluated by both UMN and PETS 2009 dataset and comparisons with the state-of-the-art method validate the advantages of our algorithm.