• Journal of the Korea Institute of Military Science and Technology
• /
• v.25 no.3
• /
• pp.300-310
• /
• 2022

A vast amount of labeled data is required for deep neural network training. A typical strategy to improve the performance of a neural network given a training data set is to use data augmentation technique. The goal of this work is to offer a novel image augmentation method for improving object detection accuracy. An object in an image is removed, and a similar object from the training data set is placed in its area. An in-painting algorithm fills the space that is eliminated but not filled by a similar object. Our technique shows at most 2.32 percent improvements on mAP in our testing on a military vehicle dataset using the YOLOv4 object detector.

• Journal of radiological science and technology
• /
• v.22 no.1
• /
• pp.5-11
• /
• 1999

This paper reviews recent major activities on instrumentation and methodology of PET. The performance of the PET instrumentation can be expressed by four physical characteristics, 1) spatial resolution, 2) coincidence resolving time, 3) energy resolution, and 4) detection efficiency. The physical and technical aspects of PET systems are briefly discussed along with these characteristics. Toward high resolution PET the recent trend has been to design multiple rings of densely packed detector arrays with scintillators. In order to satisfy the sampling requirement in reconstruction, continuous detector units has been developed. Iterative image reconstruction algorithms have received considerable attention for improvement of both the sampling requirement and image quality toward the stationary PET. Better resolving time improves the maximum true coincidence rate, which is also increased with more detectors placed in coincidence with each other. It suggests that volume PET is promising for enhancement of detection efficiency. The scattered coincidence event rate may be reduced by using detectors with better energy resolution. The use of interplane septa, however, takes over improvement of energy resolution in 2D PET. Energy resolution becomes an important factor for image quality under the condition of septa removal such as volume PET. Toward full utilization of emitting photons, 3D reconstruction incorporating oblique rays has been studied, and volume reconstruction algorithms have been developed. Practical volume PET systems impose heavy burden not only to detector sets and coincidence circuits, but also to computers in the memory requirements and the data processing. In conclusion, there have been many ingenious methods in development of PET instrumentation, which are based on unique capability of PET. They will be expected to overcome technical limitations, and to approach the fundamental limits.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.185-189
• /
• 2004

For face recognition system, a face detector which can find exact face region from complex image is needed. Many face detection algorithms have been developed under the assumption that background of the source image is quite simple . this means that face region occupy more than a quarter of the area of the source image or the background is one-colored. Color-based face detection is fast but can't be applicable to the images of which the background color is similar to face color. And the algorithm using neural network needs so many non-face data for training and doesn't guarantee general performance. In this paper, A multi-scale, multi-face detection algorithm using PCA is suggested. This algorithm can find most multi-scaled faces contained in static images with small number of training data in reasonable time.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.24 no.4
• /
• pp.392-400
• /
• 2021

Unlike optical equipment, SAR(Synthetic Aperture Radar) has the advantage of obtaining images in all weather, and object detection in SAR images is an important issue. Generally, deep learning-based object detection was mainly performed in real-valued network using only amplitude of SAR image. Since the SAR image is complex data consist of amplitude and phase data, a complex-valued network is required. In this paper, a complex-valued ResNet network is proposed. SAR image object detection was performed by combining the ROI transformer detector specialized for aerial image detection and the proposed complex-valued ResNet. It was confirmed that higher accuracy was obtained in complex-valued network than in existing real-valued network.

• Journal of Information Display
• /
• v.1 no.1
• /
• pp.29-31
• /
• 2000

A 21.6" a-Si:H TFT array for direct conversion X-ray detector with 2480 by 3072 pixels is successfully developed. To obtain X-ray image of satisfactory quality, a novel structure with a storage electrode on BCB is proposed. The structure reduces the parasitic capacitance of data line, which is one of the main sources of signal noise. Also, the structure shows greater resistance to failure than that of the conventional design.

• Journal of information and communication convergence engineering
• /
• v.4 no.2
• /
• pp.92-96
• /
• 2006

Hardware for monitoring the water quality using water fleas is developed. Water flea is a frequently used biological sensor for monitoring the water quality. Water fleas quickly respond to the incoming toxic water by changing their activity when they are exposed. By measuring the activity of water fleas, the incoming toxic water is instantly detected. So far the measurement of activity of water fleas has been done with a system equipped with both a light source of LED and a light detector of photo transistor. Water flea itself is, however, sensitive to light resulting in incorrect response and the system has two inconvenient separate parts of the light source and the detector. This paper suggests a system using a CCD camera instead of a light source and a detector. The suggested system processes the image data from the CCD camera in real time without any delay. The developed system becomes a part of the remote water monitoring embedded system.

• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.3
• /
• pp.241-246
• /
• 2016

Nowadays many people have an interest in facial expression and the behavior of people. These are human-robot interaction (HRI) researchers utilize digital image processing, pattern recognition and machine learning for their studies. Facial feature point detector algorithms are very important for face recognition, gaze tracking, expression, and emotion recognition. In this paper, a cascade facial feature point detector is used for finding facial feature points such as the eyes, nose and mouth. However, the detector has difficulty extracting the feature points from several images, because images have different conditions such as size, color, brightness, etc. Therefore, in this paper, we propose an algorithm using a modified cascade facial feature point detector using a convolutional neural network. The structure of the convolution neural network is based on LeNet-5 of Yann LeCun. For input data of the convolutional neural network, outputs from a cascade facial feature point detector that have color and gray images were used. The images were resized to $32{\times}32$. In addition, the gray images were made into the YUV format. The gray and color images are the basis for the convolution neural network. Then, we classified about 1,200 testing images that show subjects. This research found that the proposed method is more accurate than a cascade facial feature point detector, because the algorithm provides modified results from the cascade facial feature point detector.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.29 no.11 s.242
• /
• pp.1182-1188
• /
• 2005

An x-ray PIV (Particle Image Velocimetry) technique was developed to measure quantitative information on flows inside opaque conduits and on opaque-fluid flows. At first, the developed x-ray PIV technique was applied to flow in an opaque Teflon tube. To acquire x-ray images suitable for PIV velocity field measurements, refraction-based edge enhancement mechanism was employed using detectable tracer particles. The optimal distance between with the sample and detector was experimentally determined. The resulting amassed velocity field data were in reasonable agreement with the theoretical prediction. The x-ray PIV technique was also applied to blood flow in a microchannel. The flow pattern of blood was visualifed by enhancing the diffraction/interference -bas ed characteristic s of blood cells on synchrotron x-rays without any contrast agent or tracer particles. That is, the flow-pattern image of blood was achieved by optimizing the sample (blood) to detector distance and the sample thickness. Quantitative velocity field information was obtained by applying PIV algorithm to the enhanced x-ray flow images. The measured velocity field data show a typical flow structure of flow in a macro-scale channel.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.9
• /
• pp.930-935
• /
• 2014

A Gabor filter is a linear filter used for edge detectionas frequency and orientation representations of Gabor filters are similar to those of the human visual system. In this thesis, we propose a pedestrian detection algorithm using a Gabor filter bank. In order to extract the features of the pedestrian, we use various image processing algorithms and data structure algorithms. First, color image segmentation is performed to consider the information of the RGB color space. Second, histogram equalization is performed to enhance the brightness of the input images. Third, convolution is performed between a Gabor filter bank and the enhanced images. Fourth, statistical values are calculated by using the integral image (summed area table) method. The calculated statistical values are used for the feature matrix of the pedestrian area. To evaluate the proposed algorithm, the INRIA pedestrian database and SVM (Support Vector Machine) are used, and we compare the proposed algorithm and the HOG (Histogram of Oriented Gradient) pedestrian detector, presentlyreferred to as the methodology of pedestrian detection algorithm. The experimental results show that the proposed algorithm is more accurate compared to the HOG pedestrian detector.

• Proceedings of the IEEK Conference
• /
• 2003.11a
• /
• pp.241-244
• /
• 2003

Nowadays image processing is very useful for some field of traffic applications. The one reason is we can construct the system in a low price, the other is the improvement of hardware processing power, it can be more fast to processing the data. In this study, I propose the traffic monitoring system that implement on the embedded system environment. The whole system consists of two main part, one is host controller board, the other is image processing board. The part of host controller board take charge of control the total system, interface of external environment. and OSD(On screen display). The part of image processing board takes charge of image input and output using video encoder and decoder, image classification and memory control of using FPGA, control of mouse signal. And finally, fer stable operation of host controller board, uC/OS-II operating system is ported on the board.