Sang-Woo Ban | Dongguk University in Gyeongju (original) (raw)

Papers by Sang-Woo Ban

Sensors

Bearing defects are a common problem in rotating machines and equipment that can lead to unexpect... more Bearing defects are a common problem in rotating machines and equipment that can lead to unexpected downtime, costly repairs, and even safety hazards. Diagnosing bearing defects is crucial for preventative maintenance, and deep learning models have shown promising results in this field. On the other hand, the high complexity of these models can lead to high computational and data processing costs, making their practical implementation challenging. Recent studies have focused on optimizing these models by reducing their size and complexity, but these methods often compromise classification performance. This paper proposes a new approach that reduces the dimensionality of input data and optimizes the model structure simultaneously. A much lower input data dimension than that of existing deep learning models was achieved by downsampling the vibration sensor signals used for bearing defect diagnosis and constructing spectrograms. This paper introduces a lite convolutional neural network...

Applied Sciences, 2020

In this paper, we propose a deep convolutional neural network model with transfer learning that r... more In this paper, we propose a deep convolutional neural network model with transfer learning that reflects personal preferences from inter-domain databases of images having atypical visual characteristics. The proposed model utilized three public image databases (Fashion-MNIST, Labeled Faces in the Wild [LFW], and Indoor Scene Recognition) that include images with atypical visual characteristics in order to train and infer personal visual preferences. The effectiveness of transfer learning for incremental preference learning was verified by experiments using inter-domain visual datasets with different visual characteristics. Moreover, a gradient class activation mapping (Grad-CAM) approach was applied to the proposed model, providing explanations about personal visual preference possibilities. Experiments showed that the proposed preference-learning model using transfer learning outperformed a preference model not using transfer learning. In terms of the accuracy of preference recogni...

International Journal of Bio-Science and Bio-Technology, 2016

Ventricular fibrillation is one of the most common causes of sudden cardiac arrest in adults. A l... more Ventricular fibrillation is one of the most common causes of sudden cardiac arrest in adults. A lot of research has been done on ventricular fibrillation. However, that research draws conclusions from animal studies and they could not induce artificial ventricular fibrillation in human subjects. This paper looks at whether there is a difference between ventricular fibrillations in animal models and in human models, with comparisons of ventricular fibrillation obtained from the two groups. This study compares and analyzes electrocardiography (ECG) wave forms as electrical bio-signals of the two groups, in which the histogram of gradient (HOG) and auto-associative multilayer perceptrons (AAMLP) are applied for feature extraction and pattern analysis, repsectively, of the biosignals. The characteristics of electrical signals in animal ventricular fibrillation and those in human ventricular fibrillation are conclusively similar and it is reasonable to adapt the results obtained from animal research to clinical practices.

International Journal of Bio-Science and Bio-Technology, 2016

Proper indication of emergent cardiac anomalies is essential to saving human lives. Electrocardio... more Proper indication of emergent cardiac anomalies is essential to saving human lives. Electrocardiogram (ECG) signals are mainly considered for indicating cardiac status. This work proposes a model that discriminates emergent cardiac anomalies (e.g. ventricular tachyarrhythmia, congestive heart failure, malignant ventricular ectopy, supraventricular arrhythmia) from normal cardiac status using an artificial neural network. The histogram of gradient (HOG) and principal component analysis (PCA) are applied to extract generic features of the ECG signals. Five auto-associative multilayer perceptrons (AAMLP) concatenated in a cascade manner are proposed for classification of four emergent cardiac ECG signals and normal ECG signals, which was developed for implementing a primitive prototype for a mobile bio-healthcare system. Experimental results show that the proposed model successfully classifies emergent cardiac anomalies.

This paper analyses the performance of spectrum sensing in terms of the throughput of a cognitive... more This paper analyses the performance of spectrum sensing in terms of the throughput of a cognitive radio (CR) system. Dealing with the optimization problem of spectrum sensing, this paper evaluates the throughput of a CR system by considering such situations as the penalty time of a channel search and incumbent user (IU) detection delay caused by a missed detection of an incumbent signal. Also, this paper suggests a serial channel search scheme as the search method for a vacant channel, and derives its mean channel search time by considering the penalty time due to a vacant channel search error. The numerical results suggest the optimum sensing time of the channel search process using the derived mean channel search time of a serial channel search in the case of a sensing hardware structure with single radio frequency (RF) path. It also demonstrates that the average throughput is improved by two separate RF paths in spite of the hardware complexity of an RF receiver.

This paper proposes the scheme of the out-of-band cooperative spectrum sensing in the cognitive r... more This paper proposes the scheme of the out-of-band cooperative spectrum sensing in the cognitive radio (CR) base station to be operated by the multiple spectrum bands. And it suggests signal detection results for the ATSC TV signal as an incumbent signal, and derives signal detection probability and false-alarm probability for the out-of-band cooperative spectrum sensing in the frequency selective Rayleigh fading channel. Numerical results demonstrate that sensing performance for the incumbent signal is improved by the out-of-band cooperative spectrum sensing in case the power strengths of the incumbent signal to be measured by the CR users of the multiple channels are almost similar.

Journal of Korean Sensors Society, 2010

The 2006 IEEE International Joint Conference on Neural Network Proceedings, 2006

In this paper, we propose a biologically motivated face preference selective attention system to ... more In this paper, we propose a biologically motivated face preference selective attention system to identify a face within complex natural scenes. In order to localize a face in natural scenes, we have developed a task-specific selective attention model which integrates the conventional bottom-up saliency map with punishment and rewarding functions, with top-down attention and bias signals, according to a given

Lecture Notes in Computer Science, 2005

We propose a new biologically motivated model to localize or detect faces in natural color input ... more We propose a new biologically motivated model to localize or detect faces in natural color input scene. The proposed model integrates a bottom-up selective attention model and a top-down perception model. The bottom-up selective attention model using low level features sequentially selects a candidate area which is preferentially searched for face detection. The top-down perception model consists of a face spatial invariant feature detection model using ratio template matching method with training mechanism and a face color perception model, which is to model the roles of the inferior temporal areas and the V4 area, respectively. Finally, we construct a new face detection model by integration of the bottom-up saliency map model, the face color perception model and the face spatial invariant feature detection model. Computer experimental results show that the proposed model successfully indicates faces in natural scenes.

Lecture Notes in Computer Science, 2010

In this paper, we propose a new multiple sensory fused human identification model for providing h... more In this paper, we propose a new multiple sensory fused human identification model for providing human augmented cognition. In the proposed model, both facial features and mel-frequency cepstral coefficients (MFCCs) are considered as visual features and auditory features for identifying a human, respectively. As well, an adaboosting model identifies a human using the integrated sensory features of both visual and auditory features. In the proposed model, facial form features are obtained from the principal component analysis (PCA) of a human's face area localized by an Adaboost algorithm in conjunction with a skin color preferable attention model. Moreover, MFCCs are extracted from human speech. Thus, the proposed multiple sensory integration model is aimed to enhance the performance of human identification by considering both visual and auditory complementarily working under partly distorted sensory environments. A human augmented cognition system with the proposed human identification model is implemented as a goggle type, on which it presents information such as unknown people's profile based on human identification. Experimental results show that the proposed model can plausibly conduct human identification in an indoor meeting situation.

Proceedings of the 3rd …, 2009

In this paper, we propose a mew ECG pattern classification model based on a generic feature extra... more In this paper, we propose a mew ECG pattern classification model based on a generic feature extraction method. The proposed classifier is applied for indicating supraventricual arrhythmia in order to verify the performance of the proposed approach. A generic approach based on a histogram of 1 st derivative of signals is applied for feature extraction. Principal component analysis (PCA) is considered for both reducing dimension of features and extracting more plausible features from the extracted features. A simple k-means algorithm works for ECG signal classification in feature space for discriminating abnormal ECG beats caused by supraventricular arrhythmia from normal ECG ones.

Neural Information Processing-Letters and …, 2004

Abstract—In this paper, we propose a trainable selective attention model that can inhibit an unwa... more Abstract—In this paper, we propose a trainable selective attention model that can inhibit an unwanted salient area and only focus on an interesting area in a static natural scene. The proposed model was implemented by the bottom-up saliency map model in conjunction with the modified ...

Proceedings of the 9th …, 2010

In this paper we proposed an ECG(electrocardiogram) signal monitoring system working on a ZigBee ... more In this paper we proposed an ECG(electrocardiogram) signal monitoring system working on a ZigBee based wireless sensor network. An ECG signal acquisition module is implemented on a wireless platform that can acquire heart signals from ECG sensors and do wirelessly transmit the acquired heart signals based on a ZigBee protocol. Moreover, the ECG signal acquisition module is accompanied by an ECG signal monitoring module implemented in a host PC, which analyzes transmitted ECG signals from the ECG signal acquisition module and generates monitoring signals indicating normal and abnormal states. The proposed ECG signal monitoring system operating based on wireless communication of these two modules is aimed to be developed as a personalized heart signal processing system. In order to develop such a personalized system, a generic feature extraction method and an OCSVM (one-class support vector machine) classifier are applied. A histogram technique and a principal component analysis method are considered for generating features with general characteristics by extracting initial features and refined features from input ECG signals, respectively. Moreover, OCSVM is considered for developing a personalized heart signal classifier working for discriminating abnormal heart signals from normal heart signals aimed at a personalized system operating. For performance verification of the proposed system, experiments using supraventricular arrhythmia and normal ECG signals of MIT-BIH DB are conducted. The proposed system correct classification rates of 93.3% and 92.6% for normal ECG signals and supraventricular arrythmia ECG signals, respectively. Theses experimental results shows that the proposed system outperforms compared with different approaches with other classifiers.

Neural Networks, 2008

We propose new integrated saliency map and selective motion analysis models partly inspired by a ... more We propose new integrated saliency map and selective motion analysis models partly inspired by a biological visual attention mechanism. The proposed models consider not only binocular stereopsis to identify a final attention area so that the system focuses on the closer area as in human binocular vision, based on the single eye alignment hypothesis, but also both the static and dynamic features of an input scene. Moreover, the proposed saliency map model includes an affective computing process that skips an unwanted area and pays attention to a desired area, which reflects the human preference and refusal in subsequent visual search processes. In addition, we show the effectiveness of considering the symmetry feature determined by a neural network and an independent component analysis (ICA) filter which are helpful to construct an object preferable attention model. Also, we propose a selective motion analysis model by integrating the proposed saliency map with a neural network for motion analysis. The neural network for motion analysis responds selectively to rotation, expansion, contraction and planar motion of the optical flow in a selected area. Experiments show that the proposed model can generate plausible scan paths and selective motion analysis results for natural input scenes.

Neurocomputing, 2011

This paper proposes a new affective saliency map (SM) model considering psychological distance as... more This paper proposes a new affective saliency map (SM) model considering psychological distance as well as the pop-out property based on relative spatial distribution of the primitive visual features such as intensity, edge, color, and orientation. By reflecting congruency between the spatial distance caused by spatial proximity and distal in a visual scene and psychological distance caused by the way people think about visual stimuli, the proposed SM model can produce more human-like visual selective attention than a conventional SM model based on primary visual perception. In the proposed model, a psychological distance caused by a social distance, in which a proximal entity such as friend becomes more attractive when it is located near but a distal entity such as enemy becomes more attractive when it is located far from an observer, is considered. In the experiments, two types of visual stimuli are considered, mono-stimuli and stereo-stimuli. In the case of mono-stimuli, the visual stimuli on a picture with psychological depth cues were considered. Instead, in the case of stereo-stimuli, depth perception is also considered for obtaining real spatial distance of visual target in a visual scene. In order to verify the proposed affective SM model, an eye tracking system was used to measure the visual scan path and fixation time on a specific local area while monitoring the visual scenes by human subjects. Experimental results show that the proposed model can generate plausible visual selective attention properly reflecting both psychological distance and primitive visual stimuli inducing pop-out bottom-up features.

Neural Information Processing–Letters and …, 2006

Le Dong1, Sang-Woo Ban2, Inwon Lee1and Minho Lee1 1School of Electrical Engineering and Computer ... more Le Dong1, Sang-Woo Ban2, Inwon Lee1and Minho Lee1 1School of Electrical Engineering and Computer Science, Kyungpook National University 1370 Sankyuk-Dong, Puk-Gu, Taegu 702-701, Korea 2Department of Information & Communication Engineering, Dongguk ...

Sensors

Bearing defects are a common problem in rotating machines and equipment that can lead to unexpect... more Bearing defects are a common problem in rotating machines and equipment that can lead to unexpected downtime, costly repairs, and even safety hazards. Diagnosing bearing defects is crucial for preventative maintenance, and deep learning models have shown promising results in this field. On the other hand, the high complexity of these models can lead to high computational and data processing costs, making their practical implementation challenging. Recent studies have focused on optimizing these models by reducing their size and complexity, but these methods often compromise classification performance. This paper proposes a new approach that reduces the dimensionality of input data and optimizes the model structure simultaneously. A much lower input data dimension than that of existing deep learning models was achieved by downsampling the vibration sensor signals used for bearing defect diagnosis and constructing spectrograms. This paper introduces a lite convolutional neural network...

Applied Sciences, 2020

In this paper, we propose a deep convolutional neural network model with transfer learning that r... more In this paper, we propose a deep convolutional neural network model with transfer learning that reflects personal preferences from inter-domain databases of images having atypical visual characteristics. The proposed model utilized three public image databases (Fashion-MNIST, Labeled Faces in the Wild [LFW], and Indoor Scene Recognition) that include images with atypical visual characteristics in order to train and infer personal visual preferences. The effectiveness of transfer learning for incremental preference learning was verified by experiments using inter-domain visual datasets with different visual characteristics. Moreover, a gradient class activation mapping (Grad-CAM) approach was applied to the proposed model, providing explanations about personal visual preference possibilities. Experiments showed that the proposed preference-learning model using transfer learning outperformed a preference model not using transfer learning. In terms of the accuracy of preference recogni...

International Journal of Bio-Science and Bio-Technology, 2016

Ventricular fibrillation is one of the most common causes of sudden cardiac arrest in adults. A l... more Ventricular fibrillation is one of the most common causes of sudden cardiac arrest in adults. A lot of research has been done on ventricular fibrillation. However, that research draws conclusions from animal studies and they could not induce artificial ventricular fibrillation in human subjects. This paper looks at whether there is a difference between ventricular fibrillations in animal models and in human models, with comparisons of ventricular fibrillation obtained from the two groups. This study compares and analyzes electrocardiography (ECG) wave forms as electrical bio-signals of the two groups, in which the histogram of gradient (HOG) and auto-associative multilayer perceptrons (AAMLP) are applied for feature extraction and pattern analysis, repsectively, of the biosignals. The characteristics of electrical signals in animal ventricular fibrillation and those in human ventricular fibrillation are conclusively similar and it is reasonable to adapt the results obtained from animal research to clinical practices.

International Journal of Bio-Science and Bio-Technology, 2016

Proper indication of emergent cardiac anomalies is essential to saving human lives. Electrocardio... more Proper indication of emergent cardiac anomalies is essential to saving human lives. Electrocardiogram (ECG) signals are mainly considered for indicating cardiac status. This work proposes a model that discriminates emergent cardiac anomalies (e.g. ventricular tachyarrhythmia, congestive heart failure, malignant ventricular ectopy, supraventricular arrhythmia) from normal cardiac status using an artificial neural network. The histogram of gradient (HOG) and principal component analysis (PCA) are applied to extract generic features of the ECG signals. Five auto-associative multilayer perceptrons (AAMLP) concatenated in a cascade manner are proposed for classification of four emergent cardiac ECG signals and normal ECG signals, which was developed for implementing a primitive prototype for a mobile bio-healthcare system. Experimental results show that the proposed model successfully classifies emergent cardiac anomalies.

This paper analyses the performance of spectrum sensing in terms of the throughput of a cognitive... more This paper analyses the performance of spectrum sensing in terms of the throughput of a cognitive radio (CR) system. Dealing with the optimization problem of spectrum sensing, this paper evaluates the throughput of a CR system by considering such situations as the penalty time of a channel search and incumbent user (IU) detection delay caused by a missed detection of an incumbent signal. Also, this paper suggests a serial channel search scheme as the search method for a vacant channel, and derives its mean channel search time by considering the penalty time due to a vacant channel search error. The numerical results suggest the optimum sensing time of the channel search process using the derived mean channel search time of a serial channel search in the case of a sensing hardware structure with single radio frequency (RF) path. It also demonstrates that the average throughput is improved by two separate RF paths in spite of the hardware complexity of an RF receiver.

This paper proposes the scheme of the out-of-band cooperative spectrum sensing in the cognitive r... more This paper proposes the scheme of the out-of-band cooperative spectrum sensing in the cognitive radio (CR) base station to be operated by the multiple spectrum bands. And it suggests signal detection results for the ATSC TV signal as an incumbent signal, and derives signal detection probability and false-alarm probability for the out-of-band cooperative spectrum sensing in the frequency selective Rayleigh fading channel. Numerical results demonstrate that sensing performance for the incumbent signal is improved by the out-of-band cooperative spectrum sensing in case the power strengths of the incumbent signal to be measured by the CR users of the multiple channels are almost similar.

Journal of Korean Sensors Society, 2010

The 2006 IEEE International Joint Conference on Neural Network Proceedings, 2006

In this paper, we propose a biologically motivated face preference selective attention system to ... more In this paper, we propose a biologically motivated face preference selective attention system to identify a face within complex natural scenes. In order to localize a face in natural scenes, we have developed a task-specific selective attention model which integrates the conventional bottom-up saliency map with punishment and rewarding functions, with top-down attention and bias signals, according to a given

Lecture Notes in Computer Science, 2005

We propose a new biologically motivated model to localize or detect faces in natural color input ... more We propose a new biologically motivated model to localize or detect faces in natural color input scene. The proposed model integrates a bottom-up selective attention model and a top-down perception model. The bottom-up selective attention model using low level features sequentially selects a candidate area which is preferentially searched for face detection. The top-down perception model consists of a face spatial invariant feature detection model using ratio template matching method with training mechanism and a face color perception model, which is to model the roles of the inferior temporal areas and the V4 area, respectively. Finally, we construct a new face detection model by integration of the bottom-up saliency map model, the face color perception model and the face spatial invariant feature detection model. Computer experimental results show that the proposed model successfully indicates faces in natural scenes.

Lecture Notes in Computer Science, 2010

In this paper, we propose a new multiple sensory fused human identification model for providing h... more In this paper, we propose a new multiple sensory fused human identification model for providing human augmented cognition. In the proposed model, both facial features and mel-frequency cepstral coefficients (MFCCs) are considered as visual features and auditory features for identifying a human, respectively. As well, an adaboosting model identifies a human using the integrated sensory features of both visual and auditory features. In the proposed model, facial form features are obtained from the principal component analysis (PCA) of a human's face area localized by an Adaboost algorithm in conjunction with a skin color preferable attention model. Moreover, MFCCs are extracted from human speech. Thus, the proposed multiple sensory integration model is aimed to enhance the performance of human identification by considering both visual and auditory complementarily working under partly distorted sensory environments. A human augmented cognition system with the proposed human identification model is implemented as a goggle type, on which it presents information such as unknown people's profile based on human identification. Experimental results show that the proposed model can plausibly conduct human identification in an indoor meeting situation.

Proceedings of the 3rd …, 2009

In this paper, we propose a mew ECG pattern classification model based on a generic feature extra... more In this paper, we propose a mew ECG pattern classification model based on a generic feature extraction method. The proposed classifier is applied for indicating supraventricual arrhythmia in order to verify the performance of the proposed approach. A generic approach based on a histogram of 1 st derivative of signals is applied for feature extraction. Principal component analysis (PCA) is considered for both reducing dimension of features and extracting more plausible features from the extracted features. A simple k-means algorithm works for ECG signal classification in feature space for discriminating abnormal ECG beats caused by supraventricular arrhythmia from normal ECG ones.

Neural Information Processing-Letters and …, 2004

Abstract—In this paper, we propose a trainable selective attention model that can inhibit an unwa... more Abstract—In this paper, we propose a trainable selective attention model that can inhibit an unwanted salient area and only focus on an interesting area in a static natural scene. The proposed model was implemented by the bottom-up saliency map model in conjunction with the modified ...

Proceedings of the 9th …, 2010

In this paper we proposed an ECG(electrocardiogram) signal monitoring system working on a ZigBee ... more In this paper we proposed an ECG(electrocardiogram) signal monitoring system working on a ZigBee based wireless sensor network. An ECG signal acquisition module is implemented on a wireless platform that can acquire heart signals from ECG sensors and do wirelessly transmit the acquired heart signals based on a ZigBee protocol. Moreover, the ECG signal acquisition module is accompanied by an ECG signal monitoring module implemented in a host PC, which analyzes transmitted ECG signals from the ECG signal acquisition module and generates monitoring signals indicating normal and abnormal states. The proposed ECG signal monitoring system operating based on wireless communication of these two modules is aimed to be developed as a personalized heart signal processing system. In order to develop such a personalized system, a generic feature extraction method and an OCSVM (one-class support vector machine) classifier are applied. A histogram technique and a principal component analysis method are considered for generating features with general characteristics by extracting initial features and refined features from input ECG signals, respectively. Moreover, OCSVM is considered for developing a personalized heart signal classifier working for discriminating abnormal heart signals from normal heart signals aimed at a personalized system operating. For performance verification of the proposed system, experiments using supraventricular arrhythmia and normal ECG signals of MIT-BIH DB are conducted. The proposed system correct classification rates of 93.3% and 92.6% for normal ECG signals and supraventricular arrythmia ECG signals, respectively. Theses experimental results shows that the proposed system outperforms compared with different approaches with other classifiers.

Neural Networks, 2008

We propose new integrated saliency map and selective motion analysis models partly inspired by a ... more We propose new integrated saliency map and selective motion analysis models partly inspired by a biological visual attention mechanism. The proposed models consider not only binocular stereopsis to identify a final attention area so that the system focuses on the closer area as in human binocular vision, based on the single eye alignment hypothesis, but also both the static and dynamic features of an input scene. Moreover, the proposed saliency map model includes an affective computing process that skips an unwanted area and pays attention to a desired area, which reflects the human preference and refusal in subsequent visual search processes. In addition, we show the effectiveness of considering the symmetry feature determined by a neural network and an independent component analysis (ICA) filter which are helpful to construct an object preferable attention model. Also, we propose a selective motion analysis model by integrating the proposed saliency map with a neural network for motion analysis. The neural network for motion analysis responds selectively to rotation, expansion, contraction and planar motion of the optical flow in a selected area. Experiments show that the proposed model can generate plausible scan paths and selective motion analysis results for natural input scenes.

Neurocomputing, 2011

This paper proposes a new affective saliency map (SM) model considering psychological distance as... more This paper proposes a new affective saliency map (SM) model considering psychological distance as well as the pop-out property based on relative spatial distribution of the primitive visual features such as intensity, edge, color, and orientation. By reflecting congruency between the spatial distance caused by spatial proximity and distal in a visual scene and psychological distance caused by the way people think about visual stimuli, the proposed SM model can produce more human-like visual selective attention than a conventional SM model based on primary visual perception. In the proposed model, a psychological distance caused by a social distance, in which a proximal entity such as friend becomes more attractive when it is located near but a distal entity such as enemy becomes more attractive when it is located far from an observer, is considered. In the experiments, two types of visual stimuli are considered, mono-stimuli and stereo-stimuli. In the case of mono-stimuli, the visual stimuli on a picture with psychological depth cues were considered. Instead, in the case of stereo-stimuli, depth perception is also considered for obtaining real spatial distance of visual target in a visual scene. In order to verify the proposed affective SM model, an eye tracking system was used to measure the visual scan path and fixation time on a specific local area while monitoring the visual scenes by human subjects. Experimental results show that the proposed model can generate plausible visual selective attention properly reflecting both psychological distance and primitive visual stimuli inducing pop-out bottom-up features.

Neural Information Processing–Letters and …, 2006

Le Dong1, Sang-Woo Ban2, Inwon Lee1and Minho Lee1 1School of Electrical Engineering and Computer ... more Le Dong1, Sang-Woo Ban2, Inwon Lee1and Minho Lee1 1School of Electrical Engineering and Computer Science, Kyungpook National University 1370 Sankyuk-Dong, Puk-Gu, Taegu 702-701, Korea 2Department of Information & Communication Engineering, Dongguk ...