• Journal of Korea Multimedia Society
• /
• v.22 no.4
• /
• pp.491-501
• /
• 2019

Variational methods, which cast deformation as an energy-minimization problem, are known to provide a good trade-off between practicality and speed. However, the time required to deform a fully detailed shape means that these methods are largely unsuitable for real-time applications. We simplify a 2D shape into a curve skeleton, which can be deformed much more rapidly than the original shape. The curve skeleton also provides a simplified control for the user, utilizing a small number of control handles. Our system deforms the curve skeleton using an energy-minimization method and then applies the resulting deformation to the original shape using linear blend skinning. This approach effectively reduces the size of the variational optimization problem while producing deformations of a similar quality to those obtained from full-scale nonlinear variational methods.

• Journal of Multimedia Information System
• /
• v.6 no.1
• /
• pp.1-6
• /
• 2019

In the field of industrial applications, the real-time performance of the target detection problem is very important. The most serious time consumption in the pedestrian detection process is the extraction phase of the candidate window. To accelerate the speed, in this paper, a fast extraction of pedestrian candidate window based on the BING (Binarized Normed Gradients) algorithm replaces the traditional sliding window scanning. The BING features are extracted with the positive and negative samples and input into the two-stage SVM (Support Vector Machine) classifier for training. The obtained BING template may include a pedestrian candidate window. The trained template is loaded during detection, and the extracted candidate windows are input into the classifier. The experimental results show that the proposed method can extract fewer candidate window and has a higher recall rate with more rapid speed than the traditional sliding window detection method, so the method improves the detection speed while maintaining the detection accuracy. In addition, the real-time requirement is satisfied.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.4 no.2
• /
• pp.369-378
• /
• 2014

The Reed-Muller code is one of the efficient algorithms for multiple bit error correction, however, its high-computation requirement inherent in the decoding process prohibits its use in practical applications. To solve this problem, this paper proposes a graphics processing unit (GPU)-based parallel error control approach using Reed-Muller R(r, m) coding for real-time wireless communication systems. GPU offers a high-throughput parallel computing platform that can achieve the desired high-performance decoding by exploiting massive parallelism inherent in the algorithm. In addition, we compare the performance of the GPU-based approach with the equivalent sequential approach that runs on the traditional CPU. The experimental results indicate that the proposed GPU-based approach exceedingly outperforms the sequential approach in terms of execution time, yielding over 70× speedup.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.14 no.4
• /
• pp.141-147
• /
• 2014

With the rapid development of major functions in a smartphone, many applications interacting with a smartphone are introduced and mobile applications converged with various related technologies are announced and rapidly disseminated recently. In this paper, we propose the design of a smart mobile application operated with devices like sensors in order to utilize various functions of the smartphone effectively. The mobile application proposed in this paper uses the bluetooth module of the smartphone, composes the wireless ad hoc network with various sensors around us, and collects bionic context information of the user in real time. The smart application designed and implemented in this paper is the sensor based healthy diet application that can fulfill healthy diet program by reflecting the user's sensed bionic context data in real time.

• The Journal of Korea Robotics Society
• /
• v.12 no.3
• /
• pp.313-321
• /
• 2017

One of the most frequently performed tasks in human-robot interaction (HRI), intelligent vehicles, and security systems is face related applications such as face recognition, facial expression recognition, driver state monitoring, and gaze estimation. In these applications, accurate head pose estimation is an important issue. However, conventional methods have been lacking in accuracy, robustness or processing speed in practical use. In this paper, we propose a novel method for estimating head pose with a monocular camera. The proposed algorithm is based on a deep neural network for multi-task learning using a small grayscale image. This network jointly detects multi-view faces and estimates head pose in hard environmental conditions such as illumination change and large pose change. The proposed framework quantitatively and qualitatively outperforms the state-of-the-art method with an average head pose mean error of less than $4.5^{\circ}$ in real-time.

• The KIPS Transactions:PartA
• /
• v.15A no.5
• /
• pp.269-274
• /
• 2008

Recently, lots of embedded processors which can run streaming multimedia with high resolution display are introduced. Among the applications running on these embedded processors, real-time audio streaming is one of the applications that suffer from the lack of energy and memory space. In this paper, we propose a novel data compression method on scratch-pad memory, which saves both useful space on the scratch-pad memory and energy. We have implemented the data compression scheme on the GDM1202 real-time audio streaming processor, and the performance results show that we obtained 13.3% energy saving while maintaining comparable application performance to that of the non-compression case.

• Journal of Korea Multimedia Society
• /
• v.16 no.12
• /
• pp.1393-1402
• /
• 2013

Hand gesture is one of the most popular communication methods in everyday life. In human-computer interaction applications, hand gesture recognition provides a natural way of communication between humans and computers. There are mainly two methods of hand gesture recognition: glove-based method and vision-based method. In this paper, we propose a vision-based hand gesture recognition method using Kinect. By using the depth information is efficient and robust to achieve the hand detection process. The finger labeling makes the system achieve pose classification according to the finger name and the relationship between each fingers. It also make the classification more effective and accutate. Two kinds of gesture sets can be recognized by our system. According to the experiment, the average accuracy of American Sign Language(ASL) number gesture set is 94.33%, and that of general gestures set is 95.01%. Since our system runs in real-time and has a high recognition rate, we can embed it into various applications.

• Journal of Digital Contents Society
• /
• v.18 no.1
• /
• pp.115-121
• /
• 2017

The Korea metropolitan areas have the largest of population density in OECD countries. Those people who use public transportation is facing many difficulties in commuting and going to school. There are many applications to help this situation. However, those applications inform a regular timetable and also offer a real-time alarm to user by user's passive action. In this paper, we design and implement the application that give the real-time traffic information and the arrival notification message about subway, bus and foot to user.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.10 no.4
• /
• pp.7-15
• /
• 2010

Recently, there are many home network applications using the convergence between mobile platforms and home network system. As smart phones which are called small PC in the hand becoming popular, the development of home network application beyond the one of cellular phones is required. Android mobile platform of Google includes OS, middleware and primary applications. Also, in addition to smart phones, it can be mounted on various devices such as set-top box and household appliances. In this paper, home network services on smart phones of Android platform are designed and implemented. The services provide real-time monitoring information (temperature, humidity, real-time video) of rooms and appliance control in a house.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.3
• /
• pp.176-184
• /
• 1998

We present a sequential factorization method using singular value decomposition (SVD) for recovering both the three-dimensional shape of an object and the motion of camera from a sequence of images. We employ paraperpective projection [6] for camera model to handle significant translational motion toward the camera or across the image. The proposed mthod not only quickly gives robust and accurate results, but also provides results at each frame becauseit is a sequential method. These properties make our method practically applicable to real time applications. Considerable research has been devoted to the problem of recovering motion and shape of object from image [2] [3] [4] [5] [6] [7] [8] [9]. Among many different approaches, we adopt a factorization method using SVD because of its robustness and computational efficiency. The factorization method based on batch-type computation, originally proposed by Tomasi and Kanade [1] proposed the feature trajectory information using singular value decomposition (SVD). Morita and Kanade [10] have extenened [1] to asequential type solution. However, Both methods used an orthographic projection and they cannot be applied to image sequences containing significant translational motion toward the camera or across the image. Poleman and Kanade [11] have developed a batch-type factorization method using paraperspective camera model is a sueful technique, the method cannot be employed for real-time applications because it is based on batch-type computation. This work presents a sequential factorization methodusing SVD for paraperspective projection. Initial experimental results show that the performance of our method is almost equivalent to that of [11] although it is sequential.