• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.4
• /
• pp.275-285
• /
• 2003

The now video coding standard H.264 employs variable block size motion compensation, multiple references, and quarter-pel motion vector accuracy. These techniques are key features to accomplish higher coding gain, however, at the same time main factors that increase overall computational complexity. Therefore, in order to apply H.264 to many applications, key techniques are requested to improve their speed. For this reason, we propose a fast motion estimation which is suited for variable block size motion communication. In addition, we propose a fast mode decision method to choose the best mode at early stage. Experimental results show the reduction of the number of SAT SATD calculations by a factor of 4.5 and 2.6 times respectively, when we compare the proposed fast motion estimation and the conventional MVFAS $T^{［8-10］}$. Besides, the number of RDcost computations is reduced by about 45%. Therefore, the proposed methods reduces significantly its computational complexity without noticeable coding loss.

• Structural Engineering and Mechanics
• /
• v.57 no.2
• /
• pp.327-355
• /
• 2016

A new analytical derivation of the elastodynamic point load solutions for an isotropic multi-layered half-space is presented by means of the precise integration method (PIM) and the approach of dual vector. The time-harmonic external load is prescribed either on the external boundary or in the interior of the solid medium. Starting with the axisymmetric governing motion equations in a cylindrical coordinate system, a second order ordinary differential matrix equation can be gained by making use of the Hankel integral transform. Employing the technique of dual vector, the second order ordinary differential matrix equation can be simplified into a first-order one. The approach of PIM is implemented to obtain the solutions of the ordinary differential matrix equation in the Hankel integral transform domain. The PIM is a highly accurate algorithm to solve sets of first-order ordinary differential equations and any desired accuracy of the dynamic point load solutions can be achieved. The numerical simulation is based on algebraic matrix operation. As a result, the computational effort is reduced to a great extent and the computation is unconditionally stable. Selected numerical trials are given to validate the accuracy and applicability of the proposed approach. More examples are discussed to portray the dependence of the load-displacement response on the isotropic parameters of the multi-layered media, the depth of external load and the frequency of excitation.

• ETRI Journal
• /
• v.37 no.4
• /
• pp.772-779
• /
• 2015

This paper presents a power-efficient hardware realization for a motion estimation technique that is based on the full-search block matching algorithm (FSBMA). The considered input is the quarter common intermediate format of digital video. The mean of absolute difference (MAD) is the distortion criteria employed for the block matching process. The conventional architecture considered for the hardware realization of FSBMA is that of the shift register-based 2-D systolic array. For this architecture, a conservative approximation technique is adapted to eliminate unnecessary MAD computations involved in the block matching process. Upon introducing the technique to the conventional architecture, the power and complexity of its implantation is reduced, while the accuracy of the motion vector extracted from the block matching process is preserved. The proposed architecture is verified for its functional specifications. A performance evaluation of the proposed architecture is carried out using parameters such as power, area, operating frequency, and efficiency.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06d
• /
• pp.444-448
• /
• 2008

The ubiquitous smart home is the home of the future that takes advantage of context information from the human and the home environment and provides an automatic home service for the human. Human location and motion are the most important contexts in the ubiquitous smart home. We present a real-time human tracker that predicts human location and motion for the ubiquitous smart home. We used four network cameras for real-time human tracking. This paper explains the real-time human tracker's architecture, and presents an algorithm with the details of two functions (prediction of human location and motion) in the real-time human tracker. The human location uses three kinds of background images (IMAGE1: empty room image, IMAGE2:image with furniture and home appliances in the home, IMAGE3: image with IMAGE2 and the human). The real-time human tracker decides whether the human is included with which furniture (or home appliance) through an analysis of three images, and predicts human motion using a support vector machine. A performance experiment of the human's location, which uses three images, took an average of 0.037 seconds. The SVM's feature of human's motion recognition is decided from pixel number by array line of the moving object. We evaluated each motion 1000 times. The average accuracy of all the motions was found to be 86.5%.

• The KIPS Transactions:PartA
• /
• v.16A no.3
• /
• pp.209-216
• /
• 2009

The ubiquitous smart home is the home of the future that takes advantage of context information from the human and the home environment and provides an automatic home service for the human. Human location and motion are the most important contexts in the ubiquitous smart home. We present a real-time human tracker that predicts human location and motion for the ubiquitous smart home. We used four network cameras for real-time human tracking. This paper explains the real-time human tracker's architecture, and presents an algorithm with the details of two functions (prediction of human location and motion) in the real-time human tracker. The human location uses three kinds of background images (IMAGE1: empty room image, IMAGE2: image with furniture and home appliances in the home, IMAGE3: image with IMAGE2 and the human). The real-time human tracker decides whether the human is included with which furniture (or home appliance) through an analysis of three images, and predicts human motion using a support vector machine. A performance experiment of the human's location, which uses three images, took an average of 0.037 seconds. The SVM's feature of human's motion recognition is decided from pixel number by array line of the moving object. We evaluated each motion 1000 times. The average accuracy of all the motions was found to be 86.5%.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.39 no.3
• /
• pp.78-87
• /
• 2002

A macroblock(MB) in MPEG coded domain can have zero, one, or two motion vectors depending on its frame type and prediction direction (forward-, backward-, or hi-directionally). In this paper, we propose a method that converts these motion vectors on MPEG coded domain as a uniform set, independent of the frame type and the direction of prediction, and directly utilizes these re-analyzed motion vectors for understanding video contents. Also, using this frame-type-independent motion vector, we propose novel methods for detecting and tracking moving objects with frame-based detection accuracy on the compressed domain. These algorithms are performed directly from the MPEG bitstreams after VLC decoding with little time consumption. Experimental results show validity and outstanding performance of our methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.1
• /
• pp.321-340
• /
• 2016

The Compressive Video Sensing (CVS) is a useful technology for wireless systems requiring simple encoders but handling more complex decoders, and its rate-distortion performance is highly affected by the quantization of measurements and reconstruction of video frame, which motivates us to presents the Space-Time Quantization (ST-Q) and Motion-Aligned Reconstruction (MA-R) in this paper to both improve the performance of CVS system. The ST-Q removes the space-time redundancy in the measurement vector to reduce the amount of bits required to encode the video frame, and it also guarantees a low quantization error due to the fact that the high frequency of small values close to zero in the predictive residuals limits the intensity of quantizing noise. The MA-R constructs the Multi-Hypothesis (MH) matrix by selecting the temporal neighbors along the motion trajectory of current to-be-reconstructed block to improve the accuracy of prediction, and besides it reduces the computational complexity of motion estimation by the extraction of static area and 3-D Recursive Search (3DRS). Extensive experiments validate that the significant improvements is achieved by ST-Q in the rate-distortion as compared with the existing quantization methods, and the MA-R improves both the objective and the subjective quality of the reconstructed video frame. Combined with ST-Q and MA-R, the CVS system obtains a significant rate-distortion performance gain when compared with the existing CS-based video codecs.

• The Journal of the Korea Contents Association
• /
• v.16 no.2
• /
• pp.310-319
• /
• 2016

This paper proposes an effective FRUC (Frame Rate Up-Conversion) technique, which is based on ROI (Region Of Interest) separations and adaptive motion vector refinement. In this paper, in order to overcome the weakness of the EBME (Extended Bi-lateral Motion Estimation) algorithm, which is widely known in FRUC techniques, first, the proposed algorithm performs a bi-directional motion estimation for the complementary asymmetric region. Then, the proposed algorithm classifies each block into ROI or non-ROI block and refine motion vectors in accordance with their block characteristics to have a higher accuracy than the conventional EBME algorithm, specially, for the occlusion regions. The experimental results show that the proposed algorithm can improves 0.59dB on average PSNR as compared to the conventional method.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.45 no.3
• /
• pp.74-79
• /
• 2008

In an H.264/AVC video encoder, the motion estimation at fractional pixel accuracy improves a coding efficiency and image quality. However, it requires additional computation overheads for fractional search and interpolation, and thus, reducing the computation complexity of fractional search becomes more important. This paper proposes fast fractional search algorithms by combining the SASR(Simplified Adaptive Search Range) and the MSDSP(Mixed Small Diamond Search Pattern) with the predicted fractional motion vector. Compared with the full search and the prediction-based directional fractional pixel search, the proposed algorithms can reduce up to 93.2% and 81% of fractional search points, respectively with the maximum PSNR lost less than 0.04dB. Therefore, the proposed fast search algorithms are quite suitable for mobile applications requiring low power and complexity.

• Speech Sciences
• /
• v.11 no.2
• /
• pp.193-210
• /
• 2004

In this paper, we develop a real time lip-synch system that activates 2-D avatar's lip motion in synch with an incoming speech utterance. To realize the 'real time' operation of the system, we contain the processing time by invoking merge and split procedures performing coarse-to-fine phoneme classification. At each stage of phoneme classification, we apply the support vector machine (SVM) to reduce the computational load while retraining the desired accuracy. The coarse-to-fine phoneme classification is accomplished via two stages of feature extraction: first, each speech frame is acoustically analyzed for 3 classes of lip opening using Mel Frequency Cepstral Coefficients (MFCC) as a feature; secondly, each frame is further refined in classification for detailed lip shape using formant information. We implemented the system with 2-D lip animation that shows the effectiveness of the proposed two-stage procedure in accomplishing a real-time lip-synch task. It was observed that the method of using phoneme merging and SVM achieved about twice faster speed in recognition than the method employing the Hidden Markov Model (HMM). A typical latency time per a single frame observed for our method was in the order of 18.22 milliseconds while an HMM method applied under identical conditions resulted about 30.67 milliseconds.