• The KIPS Transactions:PartC
• /
• v.8C no.6
• /
• pp.805-814
• /
• 2001

In this paper, we propose a new scheduling algorithm that guarantees the delay property of real-time traffic, not considered in previous DWRR(Dynamic Weighted Round Robin) algorithm and also transmits non-real-time traffic efficiently. The proposed scheduling algorithm is a variation of DWRR algorithm to guarantee the delay property of real-time traffic by adding cell transmission method based on delay priority. It also uses the threshold to prevent the cell loss of non-real-time traffic due to cell transmission method based on delay priority. Proposed scheduling algorithm may increase some complexity over conventional DWRR scheme because of cell transmission method based on delay priority. However, the consideration of delay priority can minimize cell delay and require less size of temporary buffer. Also, the results of our performance study shows that the proposed scheduling algorithm has better performance than conventional DWRR scheme due to reliable ABR service and congestion avoidance capacity.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1993.06a
• /
• pp.229-232
• /
• 1993

이 논문은 최근에 개발된 실시간 음성부 검출 알고리즘[1]을 TMS320C30 System board와 IBM PC486을 이용한 implementation에 관한 논문이다. 음성부 검출 알고리즘은 Energy와 LCR(Level Crossing Rate)를 이용하여 각 frame을 음성/묵음으로 분류하는 방법을 사용하였고 DSP 보드를 사용하여 한 frame이 입력되면 다음 frame이 입력되기 전에 그 frame에 대한 음성/묵음 분류를 하여 음성입력이 끝남과 동시에 음성이라고 판단되는 부분만을 DPS moemory상에 저장하므로 불필요한 memory의 낭비를 중이고 다음 단계의 음성처리를 위한 시간을 절약하였다. 이 알고리즘의 성능 평가를 위하여 Rabiner와 Sambur의 알고리즘과 한민수의 알고리즘과를 전문가가 수작업으로 찾아낸 결과와 비교 평가하였다. 알고리즘의 오차는 평균 남성 4.925ms, 여성 5.85ms로 1 frame 이내의 오차를 보였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2013.11a
• /
• pp.43-45
• /
• 2013

본 논문은 안경착용 얼굴영상을 위한 실시간 눈 인식 알고리즘을 제안한다. 학습 알고리즘에 기반한 기본의 눈 인식 방법은 안경을 착용한 얼굴영상이 입력으로 주어질 경우 안경의 다양한 크기와 색깔, 형태로 인해 알고리즘의 학습 효율이 크게 떨어져 낮은 눈 인식 성능을 갖게 된다. 제안하는 방법은 모폴로지 연산을 통해 얼굴영상에서 안경이 포함된 부분을 검출하여, 안경으로 인한 눈 인식 알고리즘의 성능저하를 막는다. 성능평가를 위해 제안하는 방법을 Viola & Jones의 눈 인식 학습 기반 눈 인식 알고리즘에 적용하였으며 Spacek의 얼굴영상 데이터베이스를 실험 영상으로 사용하였다. 실험 결과, 제안하는 방법은 기존 눈 인식 알고리즘의 처리속도를 15fps (frames per second)에서 14.2fps로 하락시키면서 인식률을 75%에서 96.3%로 향상하였다.

• Journal of KIISE:Software and Applications
• /
• v.31 no.6
• /
• pp.773-783
• /
• 2004

This paper presents a method for real-time verification of Statecharts. Statecharts has been widely used for real-time reactive systems, and supports two time models: synchronous and asynchronous. However, existing real-time verification methods for them are incompatible with the asynchronous time model or increase state space by introducing new variables to the target models. We solved these problems by extending existing model checking algorithms. The extended algorithms can be used with both time models of Statecharts because they consider time increasing transitions only. In addition, they do not increase target state space since they count those transitions internally without additional variables. We extended an existing model checker, NuSMV, based on the proposed algorithms and conducted some experiments to show their advantage.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.3
• /
• pp.709-719
• /
• 1997

Performance of an algorithm can be mesaurde from serval aspects.Suppose thre is a query formulation al-gorithm.Even though this algorithm shows high retrival performance, ie, high recall and percision, retriveing items can rake a long time.In this study, we time complexity of automatic query reformulation algorithms, named the query Tree, DNF method, and Dillon's method, and comparethem in theoretical and practical aspects using a tral-time performance)the absolute times for each algorithm to fromulate a query)in a Sun SparcStation 2. In experiments using three test sets, CSCM, CISI, and Medlars, the query Tree algorithm was the fastest among the three algorithms tested.

• Journal of the Korea Computer Industry Society
• /
• v.3 no.3
• /
• pp.321-330
• /
• 2002

A 2-D median filter has many applications in various image and video signal processing areas. The rapid development in VLSI technology makes it possible to implement a real-time or near real-time 2-D median filter with reasonable cost. For the efficient VLSI implementation, the algorithm should have characteristics such as small memory requirements, regular computations, and local data transfers. This paper presents an architecture of the real-time two-dimensional separable median filter which has appropriate characteristics for the VLSI implementation. For the efficient two-dimensional median filter, a separable two-dimensional median filtering structure and a bit-sliced pipelined median searching algorithm are used. A behavioral simulator is implemented with C language and used for the analysis of the presented architecture.

• Journal of Broadcast Engineering
• /
• v.19 no.3
• /
• pp.329-341
• /
• 2014

Kernel-based mean-shift object tracking has gained more interests nowadays, with the aid of its feasibility of reliable real-time implementation of object tracking. This algorithm calculates the best mean-shift vector based on the color histogram similarity between target model and target candidate models, where the color histograms are usually produced after uniform color-space quantization for the implementation of real-time tracker. However, when the image of target model has a reduced contrast, such uniform quantization produces the histogram model having large values only for a few histogram bins, resulting in a reduced accuracy of similarity comparison. To solve this problem, a non-uniform quantization algorithm has been proposed, but it is hard to apply to real-time tracking applications due to its high complexity. Therefore, this paper proposes a fast non-uniform color-space quantization method using the histogram equalization, providing an adjusted histogram distribution such that the bins of target model histogram have as many meaningful values as possible. Using the proposed method, the number of bins involved in similarity comparison has been increased, resulting in an enhanced accuracy of the proposed mean-shift tracker. Simulations with various test videos demonstrate the proposed algorithm provides similar or better tracking results to the previous non-uniform quantization scheme with significantly reduced computation complexity.

• Journal of KIISE:Computer Systems and Theory
• /
• v.30 no.2
• /
• pp.68-77
• /
• 2003

The Least-Laxity First(or LLF) scheduling algorithm assigns the highest priority to a task with the least laxity, and has been proved to be optimal for a uni-processor and sub-optimal for a multi-processor. However, this algorithm Is Impractical to implement because laxity tie results in the frequent context switches among tasks. In this paper, a Modified Least-Laxity First on Multiprocessor(or MLLF/MP) scheduling algorithm is proposed to solve this problem, i.e., laxity tie results in the excessive scheduling overheads. The MLLF/MP is based on the LLF, but allows the laxity inversion. MLLF/MP continues executing the current running task as far as other tasks do not miss their deadlines. Consequently, it avoids the frequent context switches. We prove that the MLLF/MP is also sub-optimal in multiprocessor systems. By simulation results, we show that the MLLF/MP has less scheduling overheads than LLF.

• Annual Conference of KIPS
• /
• 2017.04a
• /
• pp.1020-1023
• /
• 2017

Viola와 Jones가 제안한 AdaBoost(Adaptive Boosting) 알고리즘은 기존의 물체 검출기에 비해 속도와 정확도 면에서 우수하여 실시간 물체 검출기로써 좋은 성능을 보인다. 하지만 여전히 많은 계산량 때문에 성능이 낮은 임베디드 환경에서는 실시간 검출에 대한 아쉬움이 있다. 본 논문에서는 계산량을 줄이기 위해 모션 검출을 통해 배경 영역을 제거하고 얼굴 영역을 추정한다. 제거된 배경 영역은 AdaBoost 알고리즘의 검출 과정에서 제외되며 추정된 얼굴 영역에 대해서만 검출을 하게 된다. 모션검출은 ${\Sigma}-{\Delta}$(Sigma-Delta) 배경 추정에 기반한 알고리즘을 사용한다.

• Annual Conference of KIPS
• /
• 2011.11a
• /
• pp.322-325
• /
• 2011

스테레오 비전 시스템에서 전통적인 매칭 알고리즘으로 SAD(Sum of Absolute Differences), SSD(Sum of Squared Differences), NCC(Normalized Cross Correlation) 등 다양한 알고리즘이 존재한다. 그러나 하드웨어로 실시간 처리를 위한 시스템을 구현하기 위해서는 리소스가 한정 되어있다는 제약 때문에 많은 연구에서 SAD 혹은 RT(Rank Transform), CT(Census Transform)를 많이 사용하게 된다. FPGA 내부에는 BRAM(Block RAM)과 MAC(multiply-accumulator)인 DSP슬라이스가 이미 존재한다. 본 논문에서는 BRAM과 DSP로직을 활용해서 전통적인 매칭 알고리즘 중에서 연산기 사용이 가장 많은 NCC를 FPGA로 실시간 처리 가능한 하드웨어 구조를 제안한다.