• Journal of Internet Computing and Services
• /
• v.5 no.6
• /
• pp.59-66
• /
• 2004

In this paper, we propose a transmission rate prediction method of video data. The proposed method checks shot transition characteristics after dividing MPEG video data into a GoP unit and then uses Kalman filter. It used algorithm to detect shot transition information by high speed in compressed domain in order to check a correct shot transition of video data and classified into a abrupt shot transition type and a gradual shot transition type. Information to have been classifying is used as factors of Kalman filter and predicts a transmission rate of video data. Also, the proposed method decreased processing time with detecting shot transition and predicting a transmission rate of video data in compressed domain. It predicted a transmission rate with 96.2- 97.6% in the experiment that used three kinds of 911 1frames of different video data.

• Economic and Environmental Geology
• /
• v.43 no.4
• /
• pp.349-358
• /
• 2010

For imaging complex subsurface structures such as salt dome, faults, thrust belt, and folds, seismic prestack reverse-time migration in depth domain is widely used, which is performed by the cross-correlation of shot-domain wavefield extrapolation with receiver-domain wavefield extrapolation. We apply the prestack reverse-time migration, which had been developed at KIGAM, to the seismic field data set of Block 5 in Jeju basin of Korea continental shelf in order to improve subsurface syncline stratigraphy image of the deep structures under the shot point 8km at the surface. We performed basic data processing for improving S/N ratio in the shot gathers, and constructed a velocity model from stack velocity which was calculated by the iterative velocity spectrum. The syncline structure of the stack image appears as disconnected interfaces due to the diffractions, but the result of the prestack migration shows that the syncline image is improved as seismic energy is concentrated on the geological interfaces.

• Journal of the Institute of Convergence Signal Processing
• /
• v.1 no.1
• /
• pp.58-63
• /
• 2000

This paper describes an efficient algorithm for shot boundary detection in MPEG video stream. There are two types of shot boundary: abrupt and gradual. The proposed algorithm for detecting the abrupt shot boundaries used DCT DC value in compressed domain. The proposed algorithm of the gradual change detection consists of two-pass mechanism. In the first pass, the expected positions of shot boundaries are extracted using ratio value of motion vectors. After decoding frames that are extracted in the first pass, we will make the dissolving image using (n)th and (n+2)th image of expected position. The gradual shot boundaries are selected by similarity of the dissolving image and the image of (n+1)th expected position. As applying the algorithm for detecting shot boundaries, the gradual changes as well as the abrupt changes are detected efficiently. Experimental results indicate that the proposed method is computationally fast for detecting shot boundaries and robust to the variation of the video characteristic that is different for the kind of videos.

• Journal of Korea Multimedia Society
• /
• v.25 no.5
• /
• pp.685-694
• /
• 2022

The goal of few-shot semantic segmentation is to build a network that quickly adapts to novel classes with extreme data shortage regimes. Most existing few-shot segmentation methods leverage single or multiple prototypes from extracted support features. Although there have been promising results for natural images, these methods are not directly applicable to the aerial image domain. A key factor in few-shot segmentation on aerial images is to effectively exploit information that is robust against extreme changes in background and object scales. In this paper, we propose a Mask-Guided Attention module to extract more comprehensive support features for few-shot segmentation in aerial images. Taking advantage of the support ground-truth masks, the area correlated to the foreground object is highlighted and enables the support encoder to extract comprehensive support features with contextual information. To facilitate reproducible studies of the task of few-shot semantic segmentation in aerial images, we further present the few-shot segmentation benchmark iSAID-, which is constructed from a large-scale iSAID dataset. Extensive experimental results including comparisons with the state-of-the-art methods and ablation studies demonstrate the effectiveness of the proposed method.

• Applied Microscopy
• /
• v.48 no.4
• /
• pp.128-129
• /
• 2018

Soft ferromagnetic materials are utilized for various electromagnetic devices such as magnetic recording heads and magnetic shielding. In situ observation of magnetic microstructures and domain wall motions are prerequisite for understanding and improving their magnetic properties. In this work, by the Fresnel (out-of-focus) method of Lorentz microscopy, we observe the domain wall motions of polycrystalline Ni/Ti thin film layers triggered by single-shot laser pulse. Random motions of domain walls were visualized at every single pulse.

• International journal of advanced smart convergence
• /
• v.8 no.4
• /
• pp.75-81
• /
• 2019

Though recent advancement of deep learning methods have provided satisfactory results from large data domain, somehow yield poor performance on few-shot classification tasks. In order to train a model with strong performance, i.e. deep convolutional neural network, it depends heavily on huge dataset and the labeled classes of the dataset can be extremely humongous. The cost of human annotation and scarcity of the data among the classes have drastically limited the capability of current image classification model. On the contrary, humans are excellent in terms of learning or recognizing new unseen classes with merely small set of labeled examples. Few-shot learning aims to train a classification model with limited labeled samples to recognize new classes that have neverseen during training process. In this paper, we increase the backbone depth of the embedding network in orderto learn the variation between the intra-class. By increasing the network depth of the embedding module, we are able to achieve competitive performance due to the minimized intra-class variation.

• Proceedings of the IEEK Conference
• /
• 2002.06d
• /
• pp.101-104
• /
• 2002

This paper proposes a unique data embedding algorithm for the video sequence. It describes two processings: shot boundary detection and robust data embedding. First, for the shot boundary detection, instead of using absolute frame differences, block by block based relative frame differences are employed. Frame adaptive thresholding values are also employed for the better detection. Second, for the robust data embedding, we generate message template and then convolve and correlate it with carrier signal. And then we embed data on the time domain video sequence. By using these two methods, watermarks into randomly selected frames of shots. Watermarks are detected well even if several certain shots are damaged because we embed watermark into each shot equally.

• Journal of the Korean Society of Systems Engineering
• /
• v.16 no.1
• /
• pp.1-8
• /
• 2020

In this paper we proposed to develop the reliability growth planning for the One-shot system using the PM2-Discrete model. The PM2-Discrete is the methodology specifically developed for discrete systems and is the first quantitative method available for formulating detailed plans in the discrete usage domain. First, the parameters RG, RI, T, MS and d of the PM2-Discrete model are set. Second, the case analysis was performed on One-shot system A. Third, the input parameter values were applied to drive the R(t) equation. Finally, using RGA 11 Software, the reliability Growth Planning Curve of One-shot system A was constructed. Also, the sensitivity analyses are performed for the changes of model parameters. The results of this study can be usefully used in establishing the reliability growth planning curve of the One-shot system.

• Journal of IKEEE
• /
• v.27 no.4
• /
• pp.537-547
• /
• 2023

In many domains, lack of data inhibits adoption of advanced machine learning models. Recently, Few-Shot Learning (FSL) has been actively studied to tackle this problem. Utilizing prior knowledge obtained through observations on related domains, FSL achieved significant performance with only a few samples. In this paper, we present a survey on FSL in terms of data augmentation, embedding and metric learning, and meta-learning. In addition to interesting researches, we also introduce major benchmark datasets. FSL is widely adopted in various domains, but we focus on image analysis in this paper.

• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.3-13
• /
• 2021

We propose a moire artifacts removal algorithm for screen-shot images using multiple domain learning. First, we estimate clean preliminary images by exploiting complementary information of the moire artifacts in pixel value and frequency domains. Next, we estimate a clean edge map of the input moire image by developing a clean edge predictor. Then, we refine the pixel and frequency domain outputs to further improve the quality of the results using the estimated edge map as the guide information. Finally, the proposed algorithm obtains the final result by merging the two refined results. Experimental results on a public dataset demonstrate that the proposed algorithm outperforms conventional algorithms in quantitative and qualitative comparison.