• 한국통신학회논문지
• /
• 제21권1호
• /
• pp.17-32
• /
• 1996

In this paper, a model for two-layered video traffic is proposed. The performance analysis of the proposed model and the effects of two-layer coding scehemes in ATM networks are also studied. ATM-based networks give the possibility to support image codingat variable bit rate(VBR). Two layer coding is one of the very promising methods among many proposed methods to compensate the cell loss, the major drawback in ATM networks. From the experimental data of the 2-layer coded video traffics, it is observed that traffic patterns of base layer and enhanced layer are highly correlate to each other, when constant image quality is kept. With this observation, coded two layered video traffic can be modeled as 2-dimensional Markov chain. The model well fit the real experimental data. The model was used for the analysis of the performance of statistical multiplexer with priorites in ATM networks.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2016년도 하계학술대회
• /
• pp.260-263
• /
• 2016

Recently high dynamic range imaging technique is hot issue in computer graphic area. We present a progressive tone mapping algorithm, which is based on weighted least squares optimization framework. Our approach combines weighted leastsquaresfiltering with iCAM06, for showing more perceptual high dynamic range images in conventional display, while avoiding visual halo artifacts. We decompose high dynamic range image into base layer and detail layer. The base layer has large scale variation, it is obtained by using weighted least squares filtering, and then the base layer incorporates iCAM06 model. Then, adaptive compression on the base layer according to human visual system. Only the base layer reduces contrast, and preserving detail. The resultshows more perceptual color appearance and preserve fine detail, while avoiding common artifacts.

• 대한전자공학회논문지
• /
• 제23권3호
• /
• pp.322-326
• /
• 1986

Nonreciprocal coupling characteristics are studied in the several coupled image guide structures containing a ferrite layer. The analysis is based upon the transverse resonance and the effective dielectric constant approach. Numerical results at 35 GHz are presented, and experimental results as wel as theoretical ones at 10GHz are also discussed.

• 전자공학회논문지B
• /
• 제30B권2호
• /
• pp.69-77
• /
• 1993

Automatic image recognition system is essential for a better man-to machine interaction. Because of the noise and deformation due to the sensor operation, it is not simple to build an image recognition system even for the fixed images. In this paper neural network which has been reported to be adequate for pattern recognition task is applied to the fixed and variational(rotation, size, position variation for the fixed image)recognition with a hope that the problems of conventional pattern recognition techniques are overcome. At fixed image recognition system. ART model is trained with face images obtained by camera. When recognizing an matching score. In the test when wigilance level 0.6 - 0.8 the system has achievel 100% correct face recognition rate. In the variational image recognition system, 65 invariant moment features sets are taken from thirteen persons. 39 data are taken to train multi-layer perceptron and other 26 data used for testing. The result shows 92.5% recognition rate.

• Structural Engineering and Mechanics
• /
• 제77권1호
• /
• pp.47-56
• /
• 2021

This paper deals with damage detection using a Gapped Smoothing Method (GSM) combined with deep learning. Convolutional Neural Network (CNN) is a model of deep learning. CNN has an input layer, an output layer, and a number of hidden layers that consist of convolutional layers. The input layer is a tensor with shape (number of images) × (image width) × (image height) × (image depth). An activation function is applied each time to this tensor passing through a hidden layer and the last layer is the fully connected layer. After the fully connected layer, the output layer, which is the final layer, is predicted by CNN. In this paper, a complete machine learning system is introduced. The training data was taken from a Finite Element (FE) model. The input images are the contour plots of curvature gapped smooth damage index. A free-free beam is used as a case study. In the first step, the FE model of the beam was used to generate data. The collected data were then divided into two parts, i.e. 70% for training and 30% for validation. In the second step, the proposed CNN was trained using training data and then validated using available data. Furthermore, a vibration experiment on steel damaged beam in free-free support condition was carried out in the laboratory to test the method. A total number of 15 accelerometers were set up to measure the mode shapes and calculate the curvature gapped smooth of the damaged beam. Two scenarios were introduced with different severities of the damage. The results showed that the trained CNN was successful in detecting the location as well as the severity of the damage in the experimental damaged beam.

• 한국멀티미디어학회논문지
• /
• 제23권7호
• /
• pp.803-811
• /
• 2020

This paper proposes an image classification algorithm that transforms the number of convolution layers in the residual block of ResNet, CNN's representative method. The proposed method modified the structure of 34/50 layer of ResNet structure. First, we analyzed the performance of small and many convolution layers for the structure consisting of only shortcut and 3 × 3 convolution layers for 34 and 50 layers. And then the performance was analyzed in the case of small and many cases of convolutional layers for the bottleneck structure of 50 layers. By applying the results, the best classification method in the residual block was applied to construct a 34-layer simple structure and a 50-layer bottleneck image classification model. To evaluate the performance of the proposed image classification model, the results were analyzed by applying to the cifar10 dataset. The proposed 34-layer simple structure and 50-layer bottleneck showed improved performance over the ResNet-110 and Densnet-40 models.

• 융합신호처리학회논문지
• /
• 제4권4호
• /
• pp.7-14
• /
• 2003

EZW(Embedded Zerotree Wavelet) 알고리즘이 소개된 이래 일련의 내장형 웨이브렛 코딩 방법들이 제안되어져왔다. 이들의 하나의 공통된 특징은 EZW 알고리즘의 기본 아이디어를 근간으로 한다는 점이다. 특히 SPIHT(Set Partitioning in Hierarchy Trees) 알고리즘은 이들 중의 하나로서 산술 코더를 사용하지 않더라도 EZW와 같거나 혹은 더 나은 성능을 제공할 수 있기 때문에 널리 알려져 왔다. 본 연구에서는 내장형 영상코딩을 위한 재귀적 SPIHT(RSPIHT) 알고리즘을 제안하고 그 유효성을 실험적으로 가한다. 제안한 RSPIHT 알고리즘은 매우 단순하고 정형화된 형태를 지니면서 최악의 경우 시간복잡도 O(n)을 가진다. 실험영상들에 대해 T-layer 4 이상에서 SPIHT보다 평균 약 16.4%의 개선된 속도를 얻을 수 있었다. 압축률의 관점에서도 RSPIHT 알고리즘은 실험영상의 T-layer 7 이하에서는 SPIHT와 유사한 결과를 가지나 그보다 큰 T-layer에서는 개선된 결과를 보였다.

• 대한전자공학회논문지
• /
• 제25권4호
• /
• pp.422-429
• /
• 1988

The photoconductive multilayer of Se-As(hole blocking layer)/Se-As-Te (photoconductive layer) /Se-As (layer for supporiting hole transport)/Se-As(layer or controlling total capacitance)/Sb2S3(electron blocking layer) was fabricated and its electrical and optical properties were investigated. The photoconductive multilayer is made of evaporated a-Se as the base material, doped with As and Te to prevent the crystallization of a-Se and to enhance red sensitivity, respectively. The multilayer with good image reproducibility has the following deposition condition. The first layer has the thickness of 250\ulcornerat the deposition rate of 250\ulcornersec. The second layer has the thickness of 800\ulcornerat the deposition rate of 250\ulcornersec. The third layer has the thickness of 125\ulcornerat the deposition rate of 250\ulcornersec. The fourth layer has the thickness of 1700\ulcornerunder the Ar gas ambient of 50x10**-3torr. The image pick-up tube, employing this multilayer demonstrates the following characteristics. The photosensitivity is 0.8, the resolution limit is above 300TV line, and the decay lag is about 7%. And spectral response convers the whole visible range. Therfore the application to color TV camera is expected.

• 치과방사선
• /
• 제21권1호
• /
• pp.119-125
• /
• 1991

In the study of magnification of image in the focal trough of panoramic radiography, (Yoshida company Panoura - Eight - s), a series of 63 x-ray films were taken with the 8-19 metal pins placed in the holes of the plastic plate, measured and evaluated by 4 observes. The author analyzed the vertical and horizontal magnification rate in the corrected focal trough. Results were as follows: 1. For vertial measurements, magnification rates were minimum 10% maximum 36% and the magnification for image of medial side was larger than that of image of lateral side from image layer. 2. For horizontal measurements, magnification rates were minimum-14% maximum 46% and images of medial side from focal trough were magnified and images of lateral side from focal trough were retrenched. 3. When moved 10㎜ downward occlusal layer, interspace was somewhat narrow between the pins and upper sides of pins were horizontally magnified but images of the end parts of pins showed tapered form. 4. When moved 10㎜ downward from the occlusal layer, opposite images showed overlapping.

• International Journal of CAD/CAM
• /
• 제9권1호
• /
• pp.37-46
• /
• 2010

Existing methods for the registration of blurred images are efficient for the artificially blurred images or a planar registration, but not suitable for the naturally blurred images existing in the real image mosaic process. In this paper, we attempt to resolve this problem and propose a method for a distortion-free stitching of naturally blurred images for image mosaic. It adopts a multi-resolution and robust feature based inter-layer mosaic together. In each layer, Harris corner detector is chosen to effectively detect features and RANSAC is used to find reliable matches for further calibration as well as an initial homography as the initial motion of next layer. Simplex and subspace trust region methods are used consequently to estimate the stable focal length and rotation matrix through the transformation property of feature matches. In order to stitch multiple images together, an iterative registration strategy is also adopted to estimate the focal length of each image. Experimental results demonstrate the performance of the proposed method.