• Journal of the Institute of Convergence Signal Processing
• /
• v.22 no.3
• /
• pp.99-103
• /
• 2021

Recently, convolutional neural networks (CNNs) have been widely utilized due to their outstanding performance in various computer vision fields. However, due to their computational-intensive and high memory requirements, it is difficult to deploy CNNs on hardware platforms that have limited resources, such as mobile devices and IoT devices. To address these limitations, a neural network compression research is underway to reduce the size of neural networks while maintaining their performance. This paper proposes a CNN compression technique that dynamically adjusts the thresholds of pruning, one of the neural network compression techniques. Unlike the conventional pruning that experimentally or heuristically sets the thresholds that determine the weights to be pruned, the proposed technique can dynamically find the optimal thresholds that prevent accuracy degradation and output the light-weight neural network in less time. To validate the performance of the proposed technique, the LeNet was trained using the MNIST dataset and the light-weight LeNet could be automatically obtained 1.3 to 3 times faster without loss of accuracy.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2003.06a
• /
• pp.230-233
• /
• 2003

ATM 트래픽 관리의 목적은 호 설정시 요구하는 QoS를 충족시키는 것과 최소한의 망 자원을 이용하면서 망을 보호하는 것이다. 또한, 서로 다른 채널간 QoS을 공정하게 보장하는 것이 필요하다. 본 논문에서는 채널들간에 공정한 QoS 제공과 링크 이용률을 높이기 위하여 입력버퍼에 문턱값을 기반으로 출력 셀율을 동적으로 조정하는 새로운 TBBM(threshold based buffer management)알고리즘을 제안하였으며 문턱값 변화에 따른 성능을 분석하였다.

• Journal of Korea Multimedia Society
• /
• v.4 no.6
• /
• pp.500-509
• /
• 2001

Image compression based on Wavelet gives much better quality than JPEG based on DCT, but suffers from ringing or blurring effects around edges as the compression is increased. In this paper, we proposed enhanced image compression by pre-processing wavelet coefficients. This pre-processing is performed by making a low threshold and enhanced by zerotree scan method when subband's zerotrees are established. It might increase significants coefficient by means of modifying the threshold and reflect on the orientation of subbands. Some experimental results show our method is more efficient than the conventional methods, JPEG. And then the developed coding scheme improves the quality of images and visually shows more pleasing results for most practical images.

• Journal of the Institute of Convergence Signal Processing
• /
• v.5 no.1
• /
• pp.79-83
• /
• 2004

The purpose of ATM traffic management is to protect the network using minimum resources and guarantee the requred QoS and also it is desirable to provide fairness between input channels. In this paper, we propose the TBBM(threshold based buffer management) algorithm to improve fairness between input channels and utilization of ATM networks. TBBM algorithm controls output cell rate dynamically based on threshold. The result shows that the required bandwidth of the TBBM algorithm is 14.3% lower in audio traffic and 41.8% lower in video traffic than that of theoretically calculated equivalent capacity method. and also reveals that the TBBM algorithm provide improved CLR fairness between input channels

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06c
• /
• pp.129-132
• /
• 1994

본 논문에서는 윈도우즈용 음성인식 software "voice access"를 개발하여 연구한 실시간 음성검출 알고리즘에 관해 소개한다. 이 음성검출 알고리즘은 200 sample 단위의 프레임 에너지, 프레임 영교차율, 음성의 길이를 음성검출의 파라메타로 사용한다. 각 파라메타의 문턱값은 신호의 평균값, 잡음의 표준편차, 미디안 표준편차와 한국어의 음성적 특성을 고려하여 설정하였으며 주변의 환경에 적응해 가며 문턱값을 조정하므로 주변 잡음환경의 변화에 대해서도 강인한 음성검출 결과를 보여준다. 또한 실시간으로 음성을 검출하므로 실용성이 높다. 음성의 검출은 일반사운드 카드를 통해 16-bit의 8KHz로 샘플링된 신호를 사용한다. 음성검출을 위한 분석은 200 sample 씩 하고 100 sample 씩 overlap 하면서 수행한다. 음성검출을 위한 모든 분석은 특별한 DSP의 도움없이 486D 이상에서 실시간으로 구현했다.시간으로 구현했다.

• Electrical & Electronic Materials
• /
• v.5 no.3
• /
• pp.320-327
• /
• 1992

Twin-tub CMOS 공정에 의해 제작된 서브마이크로미터 채널길이를 갖는 n채널 및 p채널 MOSFET의 특성을 고찰하였다. n채널 및 p채널 영역에서의 불순물 프로파일과 채널 이온주입 조건에 따른 문턱전압의 의존성 및 퍼텐셜 분포를 SUPREM-II와 MINIMOS 4.0을 사용하여 시뮬레이션하였다. 문턱전압 조정을 위한 counter-doped 보론 이온주입에 의해 p채널 MOSFET는 표면에서 대략 0.15.mu.m의 깊이에서 매몰채널이 형성되었다. 각 소자의 측정 결과, 3.3[V] 구동을 위한 충분한 여유를 갖는 양호한 드레인 포화 특성과 0.2[V]이하의 문턱전압 shift를 갖는 최소화된 짧은 채널 효과, 10[V]이상의 높은 펀치쓰루 전압과 브레이크다운 전압, 낮은 subthreshold 값을 얻었다.

• The KIPS Transactions:PartB
• /
• v.17B no.5
• /
• pp.363-374
• /
• 2010

In this paper, we propose the method for extracting Homogeneity Threshold($H_T$) and for segmenting homogeneous regions by USRG(Unseeded Region Growing) with $H_T$. The $H_T$ is a criterion to distinguish homogeneity in neighbor pixels and is computed automatically from the original image by proposed method. Theoretical background for proposed method is based on the Otsu's single level threshold method. The method is used to divide a small local part of original image int o two classes and the sum($\sigma_c$) of standard deviations for the classes to satisfy special conditions for distinguishing as different regions from each other is used to compute $H_T$. To find validity for proposed method, we compare the original image with the image that is regenerated with only the segmented homogeneous regions and show up the fact that the difference between two images is not exist visually and also present the steps to regenerate the image in order the size of segmented homogeneous regions and in order the intensity that includes pixels. Also, we show up the validity of proposed method with various results that is segmented using the homogeneity thresholds($H^*_T$) that is added a coefficient ${\alpha}$ for adjusting scope of $H_T$. We expect that the proposed method can be applied in various fields such as visualization and animation of natural image, anatomy and biology and so on.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.7A
• /
• pp.548-553
• /
• 2009

The aim of 4G mobile communication systems is to connect arbitrary devices to high-speed network anytime, anywhere. Cellular system must solve the problem of inter-cell interference caused by frequency reuse for meeting needs of 4G. In this paper, to solve interference problem we propose the algorithm, the so-called Threshold-Based Region Decision (TBRD) and analyze characteristics and performance of the proposed algorithm through simulation. The proposed technique has high applicability and flexibility due to simple construction using only existing UE measurement, and achieves performance improvement through the threshold that can be easily controlled.

• Journal of IKEEE
• /
• v.24 no.1
• /
• pp.194-199
• /
• 2020

The threshold voltage roll-off for an asymmetric junctionless double gate MOSFET is analyzed according to the top and bottom gate oxide thicknesses. In the asymmetric structure, the top and bottom gate oxide thicknesses can be made differently, so that the top and bottom oxide thicknesses can be adjusted to reduce the leakage current that may occur in the top gate while keeping the threshold voltage roll-off constant. An analytical threshold voltage model is presented, and this model is in good agreement with the 2D simulation value. As a result, if the thickness of the bottom gate oxide film is decreased while maintaining a constant threshold voltage roll-off, the top gate oxide film thickness can be increased, and the leakage current that may occur in the top gate can be reduced. Especially, it is observed that the increase of the bottom gate oxide thickness does not affect the threshold voltage roll-off.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.59-62
• /
• 2000

화상회의 및 인터넷폰을 목적으로 개발된 G.723.1은 CELP계열 보코더와 같이 화자의 개성정보를 위해 피치를 전송하고 있다. 하지만 안정구간과 비 안정구간의 차이를 두지 않고 처리를 하기 때문에 비 안정구간, 특히 전이구간에서 정확한 피치검출이 이루어지지 않는 이유로 음질의 열하가 발생하게 된다. 따라서 본 논문에서는 한 프레임 구간에서 에너지의 기울기로 대략적인 피치이득을 구한 다음 안정구간일 때와 프레임 내의 에너지의 기울기가 문턱 값을 넘을 때에는 기존의 방법으로 피치를 구하고 그런지 않은 경우에는 에너지를 조정하여 피치를 다시 구하는 방법을 사용하였다. 실제 음성시료에 대해 기존의 방법과 제안한 방법을 비교하기 위해 SegSNR 과 MOS를 비교하였을 때 각각 1.302(dB)와 평균 0.045 MOS가 향상되었다.