• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.4
• /
• pp.673-677
• /
• 2003

In this study, we propose the new filter bank that is adaptive filter bank using neural networks in time domain. Also, we proposed a new filter neuron as neuron with filter window, the structure and algorithm for filter banks. The performance of neural filter banks is shown from two examples. It show characteristics the simple structure and higher speed processing than traditional methods (filter banks in frequency domain, etc.). In many applications, the proposed method will provide the high performance to features detection of signals in time domain.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.1
• /
• pp.71-76
• /
• 2023

This paper studied a method of attenuating nonlinear noise using a Wiener filter of a neural network structure in an acoustic noise attenuator. This system improves nonlinear noise attenuation performance with a deep learning algorithm using a neural network Wiener filter instead of using a conventional adaptive filter. A voice is estimated from a single input voice signal containing nonlinear noise using a 128-neuron, 8-neuron hidden layer and an error back propagation algorithm. In this study, a simulation program using the Keras library was written and a simulation was performed to verify the attenuation performance for nonlinear noise. As a result of the simulation, it can be seen that the noise attenuation performance of this system is significantly improved when the FNN filter is used instead of the Wiener filter even when nonlinear noise is included. This is because the complex structure of the FNN filter expresses any type of nonlinear characteristics well.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.5
• /
• pp.537-542
• /
• 2011

A mirror neuron is a neuron fires both when an animal acts and when the animal observes the same action performed by another. We propose a method of 3D reconstruction for occluded object motion tracking like Mirror Neuron System to fire in hidden condition. For modeling system that intention recognition through fire effect like Mirror Neuron System, we calculate depth information using stereo image from a stereo camera and reconstruct three dimension data. Movement direction of object is estimated by optical flow with three-dimensional image data created by three dimension reconstruction. For three dimension reconstruction that enables tracing occluded part, first, picture data was get by stereo camera. Result of optical flow is made be robust to noise by the kalman filter estimation algorithm. Image data is saved as history from reconstructed three dimension image through motion tracking of object. When whole or some part of object is disappeared form stereo camera by other objects, it is restored to bring image date form history of saved past image and track motion of object.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.987-994
• /
• 2020

In this paper, we studied the effect of kernel size of CNN layer on performance in acoustic noise attenuators. This system uses a deep learning algorithm using a neural network adaptive prediction filter instead of using the existing adaptive filter. Speech is estimated from a single input speech signal containing noise using a 100-neuron, 16-filter CNN filter and an error back propagation algorithm. This is to use the quasi-periodic property in the voiced sound section of the voice signal. In this study, a simulation program using Tensorflow and Keras libraries was written and a simulation was performed to verify the performance of the noise attenuator for the kernel size. As a result of the simulation, when the kernel size is about 16, the MSE and MAE values are the smallest, and when the size is smaller or larger than 16, the MSE and MAE values increase. It can be seen that in the case of an speech signal, the features can be best captured when the kernel size is about 16.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.22 no.2
• /
• pp.135-140
• /
• 2012

The mirror neuron system in the cerebrum, which are handled by visual information-based imitative learning. When we observe the observer's range of mirror neuron system, we can assume intention of performance through progress of neural activation as specific range, in include of partially hidden range. It is goal of our paper that imitative learning is applied to 3D vision-based intelligent system. We have experiment as stereo camera-based restoration about acquired 3D image our previous research Using Optical flow, unscented Kalman filter. At this point, 3D input image is sequential continuous image as including of partially hidden range. We used Hidden Markov Model to perform the intention recognition about performance as result of restoration-based hidden range. The dynamic inference function about sequential input data have compatible properties such as hand gesture recognition include of hidden range. In this paper, for proposed intention recognition, we already had a simulation about object outline and feature extraction in the previous research, we generated temporal continuous feature vector about feature extraction and when we apply to Hidden Markov Model, make a result of simulation about hand gesture classification according to intention pattern. We got the result of hand gesture classification as value of posterior probability, and proved the accuracy outstandingness through the result.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.12 no.1
• /
• pp.75-81
• /
• 2012

Road identification is the important task for extracting a road region from the high-resolution satellite images, when the road candidates is extracted by the pre-processing tasks using a binarization, noise removal, and color processing. Therefore, we propose a noble approach for identifying a road using the orientation-selective spatial filters, which is motivated by a computational model of neuron cells found in the primary visual cortex. In our approach, after the neuron cell typed spatial filters is applied to the isolated connected-labeling road candidate regions, proposed method identifies the region of perceiving the strong orientation feature with the real road region. To evaluate the effectiveness of the proposed method, the accuracy&error ratio in the confusion matrix was measured from road candidates including road and non-road class. As a result, the proposed method shows the more than 92% accuracy.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.12 no.1
• /
• pp.60-65
• /
• 2012

This paper presents a method for occluded object based motion estimation and tracking system in dynamic image sequences using particle filter with 3D reconstruction. A unique characteristic of this study is its ability to cope with partial occlusion based continuous motion estimation using particle filter inspired from the mirror neuron system in human brain. To update a prior knowledge about the shape or motion of objects, firstly, fundamental 3D reconstruction based occlusion tracing method is applied and object landmarks are determined. And optical flow based motion vector is estimated from the movement of the landmarks. When arbitrary partial occlusions are occurred, the continuous motion of the hidden parts of object can be estimated by particle filter with optical flow. The resistance of the resulting estimation to partial occlusions enables the more accurate detection and handling of more severe occlusions.

• Journal of Advanced Marine Engineering and Technology
• /
• v.21 no.1
• /
• pp.66-70
• /
• 1997

In this paper a new learning algorithm for curvature smoothing and improved generalization for multi-layer neural networks is proposed. To enhance the generalization ability a constraint term of hidden neuron activations is added to the conventional output error, which gives the curvature smoothing characteristics to multi-layer neural networks. When the total cost consisted of the output error and hidden error is minimized by gradient-descent methods, the additional descent term gives not only the Hebbian learning but also the synaptic weight decay. Therefore it incorporates error back-propagation, Hebbian, and weight decay, and additional computational requirements to the standard error back-propagation is negligible. From the computer simulation of the time series prediction with Santafe competition data it is shown that the proposed learning algorithm gives much better generalization performance.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.6
• /
• pp.1103-1108
• /
• 2023

In this paper, we propose a new algorithm for attenuating the background noises in acoustic signal. This algorithm improves the noise attenuation performance by using the FNN(: Full-connected Neural Network) deep learning algorithm instead of the existing adaptive filter after wavelet transform. After wavelet transforming the input signal for each short-time period, noise is removed from a single input audio signal containing noise by using a 1024-1024-512-neuron FNN deep learning model. This transforms the time-domain voice signal into the time-frequency domain so that the noise characteristics are well expressed, and effectively predicts voice in a noisy environment through supervised learning using the conversion parameter of the pure voice signal for the conversion parameter. In order to verify the performance of the noise reduction system proposed in this study, a simulation program using Tensorflow and Keras libraries was written and a simulation was performed. As a result of the experiment, the proposed deep learning algorithm improved Mean Square Error (MSE) by 30% compared to the case of using the existing adaptive filter and by 20% compared to the case of using the STFT(: Short-Time Fourier Transform) transform effect was obtained.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.05
• /
• pp.398-401
• /
• 1997

The EMG signal of spinal cord injured patient is very feeble because that the information from central nervous system is not sufficiently transmitted to molter neuron or muscle fiber. Therefore the observer can not observe contraction and relaxation movement of muscle from the raw EMG signal. In this paper, we propose the muscle contraction and relaxation pattern analysis method of spinal cord injured patient whose EMG signal is composed of the sum of motor unit action potential train with additive white Gaussian noise and impulsive noise. From the EMG model, we denoise impulsive noise using median filter which is a kind of nonlinear filter and the output of median filter is transformed to wavelet transform domain for denoising additive white Gaussian noise using threshold level removal technique. As a result, we can obtain the clear contraction and relaxation pattern.