• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.180-194
• /
• 2021

For small image registration, feature-based approaches are likely to fail as feature detectors cannot detect enough feature points from low-resolution images. The classic FFT approach's prediction accuracy is high, but the registration time can be relatively long, about several seconds to register one image pair. To achieve real-time and high-precision rigid registration for small images, we apply deep neural networks for supervised rigid transformation prediction, which directly predicts the transformation parameters. We train deep registration models with rigidly transformed CIFAR-10 images and STL-10 images, and evaluate the generalization ability of deep registration models with transformed CIFAR-10 images, STL-10 images, and randomly generated images. Experimental results show that the deep registration models we propose can achieve comparable accuracy to the classic FFT approach for small CIFAR-10 images (32×32) and our LSTM registration model takes less than 1ms to register one pair of images. For moderate size STL-10 images (96×96), FFT significantly outperforms deep registration models in terms of accuracy but is also considerably slower. Our results suggest that deep registration models have competitive advantages over conventional approaches, at least for small images.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.12C
• /
• pp.1167-1176
• /
• 2009

In This Paper, we introduce an efficient IFFT design technique named for transmitter of OFDM (Orthogonal Frequency Division Multiplexing) system. In OFDM system, a cyclic prefix is inserted in forepart of OFDM symbol to prevent ICI(Inter-channel Interference) and ISI (Inter-symbol Interference). Attaching cyclic prefix causes delay in storing and copying IFFT result. The proposed IFFT removes this delay because its output is cyclic shifted by the length of cyclic prefix. So we can make a complete OFDM symbol by just copying the forepart of IFFT output to the end. In many cases, the length of cyclic prefix is 1/2n of FFT size, and this IFFT does not require additional hardware complexity and it does not cause any performance degradation.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.37 no.5
• /
• pp.502-510
• /
• 2009

Laser Inertial Navigation System(LINS) consists of Ring Laser Gyroscopes(RLG) and accelerometers. RLG has a lock-in region in which there is zero output for input angular rates less than 0.1deg/sec. The lock-in region is generated by the imperfect mirrors in RLG. To avoid the lock-in region, sinusoidal motion which is called dither motion is applied on RLG. Therefore without the fault in LINS, the dither motion must be measured by RLG/accelerometer. In this paper, we propose the method to detect the fault of LINS through checking out whether or not the dither motion is always measured by RLG/accelerometer using the Fast Fourier Transformation(FFT) on the real time. The feasibility of the fault detection method proposed in this paper is verified through the stationary and van test.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.10 no.1
• /
• pp.102-110
• /
• 2017

In this paper, we have implemented an real-time adaptive feedback cancellation(AFC) algorithm that can be applied to multi-channel digital hearing aid. Multichannel digital hearing aid typically use the FFT filterbank based Wide Dynamic Range Compression(WDRC) algorithm to compensate for hearing loss. The implemented real-time acoustic feedback cancellation algorithm has one integrated structure using the same FFT filter bank with WDRC, which can be beneficial in terms of computation affecting the hearing aid battery life. In addition, when the AFC fails to operate due to nonlinear input and output, the reduction gain is applied to improve robustness in practical environment. The implemented algorithm can be further improved by adding various signal processing algorithm such as speech enhancement.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.14 no.1
• /
• pp.43-53
• /
• 1990

The manufacturing errors such as pitch error and run-out error in planetary gear system bring about the irregular displacement of the center of each gear, which cause the torqe variation, vibration and noise. In this study, the relation between manufacturing errors and error motions of the center of gear was analyzed, and it can be applied to identyfy the errors of gears by investigating the measured locus of the center of each gear. Also, another identification method of power spectrum estimation using FFT algorithm was introduced, which analyze the frequency of the measured error motions. The results show that the error of each gear had a corresponding unique frequency, therefore, this method proved to be more effective.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.1A
• /
• pp.17-24
• /
• 2012

In this paper, we described a received signal model in terms of the starting point of FFT window and derive a post-detection SINR for the receiver with MMSE filter and the corresponding filter coefficients in order to mitigate the effects of interferences caused by time and frequency selective fading channels and time difference of arrival (TDoA) in OFDMA-based Mesh Networks. In addition, we proposed an MMSE Bidirectional Successive Detection (BSD) technique which can suppress the effects of interferences among multi-nodes without any redundant FFT operations. It was shown by simulation that the proposed interference suppression technique has not an error floor at higher average SNR than 30dB in terms of 64QAM uncoded BER contrary to the conventional techniques.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.1
• /
• pp.149-157
• /
• 2008

If we use the microwave of SAR, we can observe ocean in spite of severe weather or night time. The sea surface image of SAR has numerous information about atmospheric phenomena related to surface wind field. The extracted wind information from SAR can be used diversely. In order to extract sea wind speed from SAR image, a generated wind direction from SAR and sigma nought should be input into wind model. Therefore, wind speed can be obtained by input wind direction into CMOD5 Model. Azimuth angle using CMOD5 Model is generated by added $90^{\circ}$ to Look angle which is extracted from SAR data file. A gained wind direction spectrum from SAR image has $180^{\circ}$ ambiguity because of 2D-FFT. This ambiguity should decide to use the location of land, wind direction in field or the result of numerical model. Consequently, wind direction using 2D-FFT is $3^{\circ}{\sim}7^{\circ}$ differences with actual surveying data. Wind speed by CMOD5 model is similar to actual surveying data as below 2m/s.

• Journal of the Society of Disaster Information
• /
• v.19 no.4
• /
• pp.897-907
• /
• 2023

Purpose: In this study, We aim to estimate a untrained person's three postures using a 2D CNN model which is trained with minimal FFT data collected by a 24GHz FMCW radar. Method: In an indoor space, we collected FFT data for three distinct postures (standing, sitting, and lying) from three different individuals. To apply this data to a 2D CNN model, we first converted the collected data into 2D images. These images were then trained using the 2D CNN model to recognize the distinct features of each posture. Following the training, we evaluated the model's accuracy in differentiating the posture features across various individuals. Result: According to the experimental results, the average accuracy of the proposed scheme for the three postures was shown to be a 89.99% and it outperforms the conventional 1D CNN and the SVM schemes. Conclusion: In this study, we aim to estimate any person's three postures using a 2D CNN model and a 24GHz FMCW radar for disastrous situations in indoor. it is shown that the different posture of any persons can be accurately estimated even though his or her data is not used for training the AI model.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2475-2478
• /
• 2003

CELP계열의 부호화기인 G.723.1 5.3kbps ACELP를 기반으로 하여 음질을 유지하면서 전송률을 낮출 수 있는 새로운 부호화 방법을 제안한다. 본 논문에서 적용한 부호화 방법은 음성 합성시 파라미터로 사용되는 지속시간 변경에 의해 CELP형 보코더의 전송률을 감소하고자 한다. 먼저 음성을 보코더 입력단에 입력하기 전 지속시간을 FFT 변환 특성을 이용해 음색의 변경 없이 지속시간을 줄임으써 계산시간을 줄이고 진폭과 위상 각각 1/2ⁿ배의 interpolation과 Decimation을 수행하여 부호화한다. 이렇게 부호화된 데이터는 G.723.1 복호화를 거치고, 다시 FFT point의 1/2ⁿ배 point로 IFFT과정을 수행함으로써 스팩트럼의 변경 없이 지속시간을 변경하여 원 음성을 합성하게 된다. G.723.1 보코더를 통과한 후 파형을 복원 실험한 결과 기존의 5.3kbps ACELP보다 46％정도 감소하였다.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.50 no.8
• /
• pp.351-358
• /
• 2001

This paper deals with the PSS modeling of Ulchin N/P #1 as well as the generator and excitation system modeling by utilizing the recorded data from AVR step test, which has been performed by entering small voltage signal into the AVR summing point. In addition to it. two recorded results obtained from the AVR step test with PSS sunning and without PSS running have not only been compared each other on the time domain, but also they heve been analyzed with FFT analysis on the frequency domain; thus, the desirable effects of running PSS in Ulchin N/P #1 on power system have been explicitly confirmed. Finally, the derived PSS model parameters lead to good matches between simulation results and recorded data.