• Geophysics and Geophysical Exploration
• /
• v.24 no.3
• /
• pp.89-97
• /
• 2021

Fault detection in seismic data is well suited to the application of machine learning algorithms. Accordingly, various machine learning techniques are being developed. In recent studies, machine learning models, which utilize synthetic data, are the particular focus when training with deep learning. The use of synthetic training data has many advantages; Securing massive data for training becomes easy and generating exact fault labels is possible with the help of synthetic training data. To interpret real data with the model trained by synthetic data, the synthetic data used for training should be geologically realistic. In this study, we introduce a method to generate realistic synthetic seismic data. Initially, reflectivity models are generated to include realistic fault structures, and then, a one-way wave equation is applied to efficiently generate seismic stack sections. Next, a migration algorithm is used to remove diffraction artifacts and random noise is added to mimic actual field data. A convolutional neural network model based on the U-Net structure is used to verify the generated synthetic data set. From the results of the experiment, we confirm that realistic synthetic data effectively creates a deep learning model that can be applied to field data.

• Journal of the Institute of Convergence Signal Processing
• /
• v.22 no.4
• /
• pp.185-192
• /
• 2021

Various methods have been proposed to determine the defect detection of welding bead, and recently sensor data and image data inspection have been steadily announced. There are advantages that sensor data inspection is highly accurate, and two-dimensional-based image data inspection is able to determine the position of the welding bead. However, when analyzing only with sensor data, it is difficult to determine whether the welding has been performed at the correct position. On the other hand, the image data inspection does not have high accuracy due to noise and measurement errors. In this paper, we propose a method that can complement the shortcomings of each inspection method and increase its advantages to improve accuracy and speed up inspection by fusing sensor data inspection which are average current, average volt, and mixed gas data, and image data inspection methods and is implemented as software. In addition, it is intended to allow users to conveniently and intuitively analyze and grasp the results by performing analysis using a graphical user interface(GUI) and checking the data and inspection results used for the inspection. Sensor inspection is performed using the characteristics of each sensor data, and image data is inspected by applying a morphology geodesic active contour algorithm. The experimental results showed 98% accuracy, and when performing the inspection on the four image data, and sensor data the inspection time was about 1.9 seconds, indicating the performance of software that can be used as a real-time inspector in the welding process.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.4
• /
• pp.446-460
• /
• 2022

In this paper, we propose an algorithm to improve target detection rate degradation due to direct blast in a bi-static sonar systems with high duty cycle using an adaptive filters. It is very important to suppress the direct blast in the aforementioned sonar systems because it has a fatal effect on the actual system operation. In this paper, the performance was evaluated by applying the Normalized Least Mean Square (NLMS) and Recursive Least Square (RLS) algorithms to the simulation and sea experimental data. The beam signals of the target and direct blast bearings were used as the input and desired signals, respectively. By optimizing the difference between the two signals, the direct blast is removed and only the target signal is remained. As a result of evaluating the results of the matched filter in the simulation, it was confirmed that the direct blast was removed to the noise level in both Linear Frequency Modultated (LFM) and Generalized Sinusoidal Frequency Modulated (GSFM), and in the case of GSFM, the target sidelobe decreased by more than 20 dB, thereby improving performance. In the sea experiment, it was confirmed that the LFM reduced the level of the transmitted direct wave by 10 dB, the GSFM reduced the level of the transmitted direct wave by about 4 dB, and the side lobe of the target decreased by about 4 dB, thereby improving the performance.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.40 no.11
• /
• pp.51-62
• /
• 2003

This paper presents a novel survivor memeory management and decoding techniques with sequential backward state transition control in the trace back Viterbi decoder. The Viterbi algorithm is an maximum likelihood decoding scheme to estimate the likelihood of encoder state for channel error detection and correction. This scheme is applied to a broad range of digital communication such as intersymbol interference removing and channel equalization. In order to achieve the area-efficiency VLSI chip design with high throughput in the Viterbi decoder in which recursive operation is implied, more research is required to obtain a simple systematic parallel ACS architecture and surviver memory management. As a method of solution to the problem, this paper addresses a progressive decoding algorithm with sequential backward state transition control in the trace back Viterbi decoder. Compared to the conventional trace back decoding techniques, the required total memory can be greatly reduced in the proposed method. Furthermore, the proposed method can be implemented with a simple pipelined structure with systolic array type architecture. The implementation of the peripheral logic circuit for the control of memory access is not required, and memory access bandwidth can be reduced Therefore, the proposed method has characteristics of high area-efficiency and low power consumption with high throughput. Finally, the examples of decoding results for the received data with channel noise and application result are provided to evaluate the efficiency of the proposed method.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.1
• /
• pp.81-87
• /
• 2010

The coverage ability is one of essential techniques for the Robotic Vacuum Cleaner (RVC). Most of the RVCs rely on random or regular pattern movement to cover a target space due to the technical difficulties to implement localization and map and constraints of hardwares such as controller and sensors. In this paper, we consider two main issues which are low computational load and using sensors with very limited sensing capabilities. First, in our approach, computing procedures to build map and detect the RVC's position are minimized by simplifying data obtained from sensors. To reduce computational load, it needs simply presenting an environment with objects of various shapes. Another isuue mentioned above is regarded as one of the most important problems in our approach, because we consider that many RVCs use low-cost sensor systems such as an infrared sensor or ultrasonic sensor with limited capabilities in limited range, detection uncertainty, measurement noise, etc. Methods presented in this paper are able to apply to general RVCs equipped with these sensors. By both simulation and real experiment, we evaluate our method and verify that the proposed method guarantees a complete coverage.

• Journal of KIISE:Software and Applications
• /
• v.34 no.7
• /
• pp.646-654
• /
• 2007

In recent years, the use of text inserted into TV contents has grown to provide viewers with better visual understanding. In this paper, video text is defined as superimposed text region located of the bottom of video. Video text extraction is the first step for video information retrieval and video indexing. Most of video text detection and extraction methods in the previous work are based on text color, contrast between text and background, edge, character filter, and so on. However, the video text extraction has big problems due to low resolution of video and complex background. To solve these problems, we propose a method to extract text from videos using the Harris corner detector. The proposed algorithm consists of four steps: corer map generation using the Harris corner detector, extraction of text candidates considering density of comers, text region determination using labeling, and post-processing. The proposed algorithm is language independent and can be applied to texts with various colors. Text region update between frames is also exploited to reduce the processing time. Experiments are performed on diverse videos to confirm the efficiency of the proposed method.

• Korean Journal of Remote Sensing
• /
• v.37 no.3
• /
• pp.475-491
• /
• 2021

With the recent increase in available high-resolution (< ~1 m) satellite SAR images, the demand for precise registration of SAR images is increasing in various fields including change detection. The registration between high-resolution SAR images acquired in different look angle is difficult due to speckle noise and geometric distortion caused by the characteristics of SAR images. In this study, registration is performed in two stages, coarse and fine, using the x-band SAR data imaged at staring spotlight mode of TerraSAR-X. For the coarse registration, a method combining the adaptive sampling method and SAR-SIFT (Scale Invariant Feature Transform) is applied, and three rigid methods (NCC: Normalized Cross Correlation, Phase Congruency-NCC, MI: Mutual Information) and one non-rigid (Gefolki: Geoscience extended Flow Optical Flow Lucas-Kanade Iterative), for the fine registration stage, was performed for performance comparison. The results were compared by using RMSE (Root Mean Square Error) and FSIM (Feature Similarity) index, and all rigid models showed poor results in all image combinations. It is confirmed that the rigid models have a large registration error in the rugged terrain area. As a result of applying the Gefolki algorithm, it was confirmed that the RMSE of Gefolki showed the best result as a 1~3 pixels, and the FSIM index also obtained a higher value than 0.02~0.03 compared to other rigid methods. It was confirmed that the mis-registration due to terrain effect could be sufficiently reduced by the Gefolki algorithm.

• Journal of Convergence for Information Technology
• /
• v.11 no.6
• /
• pp.33-39
• /
• 2021

Speech recognition technology is being combined with deep learning and is developing at a rapid pace. In particular, voice recognition services are connected to various devices such as artificial intelligence speakers, vehicle voice recognition, and smartphones, and voice recognition technology is being used in various places, not in specific areas of the industry. In this situation, research to meet high expectations for the technology is also being actively conducted. Among them, in the field of natural language processing (NLP), there is a need for research in the field of removing ambient noise or unnecessary voice signals that have a great influence on the speech recognition recognition rate. Many domestic and foreign companies are already using the latest AI technology for such research. Among them, research using a convolutional neural network algorithm (CNN) is being actively conducted. The purpose of this study is to determine the non-voice section from the user's speech section through the convolutional neural network. It collects the voice files (wav) of 5 speakers to generate learning data, and utilizes the convolutional neural network to determine the speech section and the non-voice section. A classification model for discriminating speech sections was created. Afterwards, an experiment was conducted to detect the non-speech section through the generated model, and as a result, an accuracy of 94% was obtained.

• Journal of Biomedical Engineering Research
• /
• v.25 no.5
• /
• pp.421-430
• /
• 2004

Clinically, it is almost impossible for a physician to distinguish subtle changes of frequency spectrum by using a stethoscope alone especially in the early stage of thrombus formation. Considering that reliability of mechanical valve is paramount because the failure might end up with patient death, early detection of valve thrombus using noninvasive technique is important. Thus the study was designed to provide a tool for early noninvasive detection of valve thrombus by observing shift of frequency spectrum of acoustic signals with computer aid diagnosis system. A thrombus model was constructed on commercialized mechanical valves using polyurethane or silicon. Polyurethane coating was made on the valve surface, and silicon coating on the sewing ring of the valve. To simulate pannus formation, which is fibrous tissue overgrowth obstructing the valve orifice, the degree of silicone coating on the sewing ring varied from 20％, 40％, 60％ of orifice obstruction. In experiment system, acoustic signals from the valve were measured using microphone and amplifier. The microphone was attached to a coupler to remove environmental noise. Acoustic signals were sampled by an AID converter, frequency spectrum was obtained by the algorithm of spectral analysis. To quantitatively distinguish the frequency peak of the normal valve from that of the thrombosed valves, analysis using a neural network was employed. A return map was applied to evaluate continuous monitoring of valve motion cycle. The in-vivo data also obtained from animals with mechanical valves in circulatory devices as well as patients with mechanical valve replacement for 1 year or longer before. Each spectrum wave showed a primary and secondary peak. The secondary peak showed changes according to the thrombus model. In the mock as well as the animal study, both spectral analysis and 3-layer neural network could differentiate the normal valves from thrombosed valves. In the human study, one of 10 patients showed shift of frequency spectrum, however the presence of valve thrombus was yet to be determined. Conclusively, acoustic signal measurement can be of suggestive as a noninvasive diagnostic tool in early detection of mechanical valve thrombosis.

• The Proceeding of the Korean Institute of Electromagnetic Engineering and Science
• /
• v.3 no.2
• /
• pp.30-38
• /
• 1992

In this paper, in order to apply the $\pi$/4 shift QPSK to TCM, we propose the $\pi$/8 shift 8PSK modulation technique and the trellis-coded $\pi$/8 shift 8PSK performing signal set expansion and set partition by phase difference. In addition, the Viterbi decoder with branch metrics of the squared Euclidean distance of the first phase difference as well as the Lth phase difference is introduced in order to improve the bit error rate(BER) performance in differential detection of the trellis-coded $\pi$/8 shift 8 PSK. The proposed Viterbi decoder is conceptually the same as the sliding multiple de- tection by using the branch metric with first and Lth order phase difference. We investigate the performance of the uncoded .pi. /4 shift QPSK and the trellis-coded $\pi$/8 shift 8PSK with or without the Lth phase difference metric in an additive white Gaussian noise (AWGN) and Rayleigh fading channel using the Monte Carlo simulation. The study shows that the $\pi$/4 shift QPSK with the Trellis-code i. e. the trellis-coded $\pi$/8 shift 8PSK is an attractive scheme for power and bandlimited systems and especially, the Viterbi decoder with first and Lth phase difference metrics improves BER performance. Also, the next proposed algorithm can be used in the TC $\pi$/8 shift 8PSK as well as TC MDPSK.