鷲沢 嘉一
| 情報・ネットワーク工学専攻 | 准教授 |
| Ⅱ類（融合系） | 准教授 |
- プロフィール:
-株式会社東芝及び東芝ソリューション株式会社にて文字認識システムの研究開発に従事
-独立行政法人理化学研究所にて、脳信号処理及び機械学習の研究に従事
研究者情報
研究キーワード
経歴
委員歴
- 2017年04月01日 - 2021年06月30日
英文論文誌編集委員, 電子情報通信学会 - 2021年01月01日
Secretary, IEEE SP Society Tokyo Joint Chapter - 2014年07月14日 - 2015年07月15日
英文論文誌特集号編集委員, 電子情報通信学会信号処理研究専門委員会 - 2014年04月01日
電子情報通信学会信号処理研究専門委員会, 電子情報通信学会信号処理研究専門委員会 - 2014年
常任査読委員, 電子情報通信学会 情報・システムソサイエティ, 学協会 - 2013年03月
APSIPA Biomedical Signal Processing and Systems Technical Committee Members, Asia Pacific Signal and Information Processing Association, 学協会
研究活動情報
論文
- Complex CNN incorporating Hilbert transform for steady-state visual evoked potential BCI
R. Takata; Y. Washizawa
ラスト(シニア)オーサー, Proc. of APSIPA Annual Summit and Conference 2024., 出版日 2024年12月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Effect of Touching Care on Fear in French and Japanese Subjects
François-Beno{^i}t Vialatte; Tsubasa Tokunaga; Yoshikazu Washizawa; Kazuko Hiyoshi
The Fifteenth International Conference on Advanced Cognitive Technologies and Applications COGNITIVE 2023, 出版日 2023年06月, 査読付 - EEG Evaluation of the Effects of Touching Intervention on Frustrating Tasks
Takashi MURAMATSU; Yoshikazu WASHIZAWA; Carl B. BECKER; Kazuko HIYOSHI
International Journal of Affective Engineering, JAPAN SOC KANSEI ENGINEERING, 21巻, 3号, 掲載ページ 151-157, 出版日 2022年07月, 査読付, Touching is a medical action wherein a practitioner touches a patient???s body. Touching has positive physical and mental effects, and is an important technique for nursing. This study examines the effect of touching on patient frustration. To induce frustration, we adopted a mouse pointer-moving game and a calculation task: the game required moving a mouse pointer from the start to the goal without touching walls or obstacles. We asked the participants??? acquaintance to gently touch their backs during the intervention trials. For evaluation, we used power spectral density (PSD) and electroencephalogram (EEG) event-related potential (ERP), and participants??? self-evaluation scores. Theta, alpha, and beta band PSDs increased in frustrating tasks compared to the resting state, however, PSD increments of touching intervention tasks were less than that of control tasks. These results confirm that an acquaintance???s touching can reduce frustration in difficult tasks, and concomitantly reduce unpleasant emotions.
研究論文（学術雑誌）, 英語 - Steady-State Visual Evoked Potential Classification Using Complex Valued Convolutional Neural Networks
Akira Ikeda; Yoshikazu Washizawa
Sensors, MDPI, 21巻, 掲載ページ 5309, 出版日 2021年08月06日, 査読付
研究論文（学術雑誌）, 英語 - EEG Analysis of Nursing Touch for Frustrating Work
Takashi Muramatsu; Yoshikazu Washizawa; Kazuko Hiyoshi
Proc. of 2020 IEEE 2nd Global Conference on Life Science and Technologies, 出版日 2020年03月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Spontaneous EEG Classification Using Complex Valued Neural Network
A. Ikeda; Y. Washizawa
26th International Conference on Neural Information Processing of the Asia-Pacific Neural Network Society (ICONOP), 0巻, 出版日 2019年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Music training is associated with magnetoencephalographic gamma- and alpha-band oscillation binding with attention and consciousness.
Y. Urakami; Y. Washizawa
Proc. of 8th International Congress of Pathophysiology (ICP 2018), 0巻, 0号, 掲載ページ 0-0, 出版日 2018年09月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Ranking deep neural network for automatic music recommendation system using EEG
H. Itoga, Y. Washizawa, and Y. Urakami, 0巻, 0号, 掲載ページ 0-0, 出版日 2018年03月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - EEG based automatic music recommendation system using ranking deep artificial neural network
H. Itoga; Y. Washizawa; Y. Urakami
Proc. 8th International Congress of Pathophysiology, 0巻, 0号, 掲載ページ 0-0, 出版日 2018年
研究論文（国際会議プロシーディングス）, 英語 - Restoration of dry electrode EEG using deep convolutional neural network
Y. Kojoma; Y. Washizawa
Proc. APSIPA Annual Summit and Conference 2018, accepted巻, 掲載ページ 0-0, 出版日 2018年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Discriminative sparse representation learning using multiclass hinge loss
R. Kamiya; Y. Washizawa
Proc. APSIPA Annual Summit and Conference 2018, accepted巻, 掲載ページ 0-0, 出版日 2018年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Asynchronous stimulation method for N100-P300 Speller
N. Morita; Y. Washizawa
in Proceedings of the 6th International Conference on Cognitive Neurodynamics (ICCN2017), 0巻, 0号, 掲載ページ 0-0, 出版日 2017年08月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - A Comparison of Pseudo Noise Coding and Mixed Frequency Phase Coding for Visual Evoked Potential Brain Computer Interface
J. Sato; Y. Washizawa
in Proceedings of the 2017 RISP International Workshop on Nonlinear Circuits, Communications and Signal Processing (NCSP), 1巻, 掲載ページ 85-88, 出版日 2017年03月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - An N100-P300 Spelling Brain-Computer Interface with Detection of Intentional Control
H. Sato; Y. Washizawa
Computers, 4巻, 4:31号, 出版日 2016年12月02日, 査読付
研究論文（学術雑誌）, 英語 - Learning Subspace Classification Using Subset Approximated Kernel Principal Component Analysis
Yoshikazu Washizawa
IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS, IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E99D巻, 5号, 掲載ページ 1353-1363, 出版日 2016年05月, 査読付, We propose a kernel-based quadratic classification method based on kernel principal component analysis (KPCA). Subspace methods have been widely used for multiclass classification problems, and they have been extended by the kernel trick. However, there are large computational complexities for the subspace methods that use the kernel trick because the problems are defined in the space spanned by all of the training samples. To reduce the computational complexity of the subspace methods for multiclass classification problems, we extend Oja's averaged learning subspace method and apply a subset approximation of KPCA. We also propose an efficient method for selecting the basis vectors for this. Due to these extensions, for many problems, our classification method exhibits a higher classification accuracy with fewer basis vectors than does the support vector machine (SVM) or conventional subspace methods.
研究論文（学術雑誌）, 英語 - Discriminative Metric Learning on Extended Grassmann Manifold for Classification of Brain Signals
Yoshikazu Washizawa
IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES, IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E99A巻, 4号, 掲載ページ 880-883, 出版日 2016年04月, 査読付, Electroencephalography (EEG) and magnetoencephalography (MEG) measure the brain signal from spatially-distributed electrodes. In order to detect event-related synchronization and desynchronization (ERS/ERD), which are utilized for brain-computer/machine interfaces (BCI/BMI), spatial filtering techniques are often used. Common spatial potential (CSP) filtering and its extensions which are the spatial filtering methods have been widely used for BCIs. CSP transforms brain signals that have a spatial and temporal index into vectors via a covariance representation. However, the variance-covariance structure is essentially different from the vector space, and not all the information can be transformed into an element of the vector structure. Grassmannian embedding methods, therefore, have been proposed to utilize the variance-covariance structure of variational patterns. In this paper, we propose a metric learning method to classify the brain signal utilizing the covariance structure. We embed the brain signal in the extended Grassmann manifold, and classify it on the manifold using the proposed metric. Due to this embedding, the pattern structure is fully utilized for the classification. We conducted an experiment using an open benchmark dataset and found that the proposed method exhibited a better performance than CSP and its extensions.
研究論文（学術雑誌）, 英語 - Vibrotactile Brain-Computer Interface with Error-Detecting Codes
Sittipong Apichartstaporn; Kitsuchart Pasupa; Yoshikazu Washizawa
ADVANCES IN COGNITIVE NEURODYNAMICS (V), SPRINGER-VERLAG SINGAPORE PTE LTD, 掲載ページ 355-361, 出版日 2016年, 査読付, Nowadays, Brain-Computer Interfaces (BCIs) have been used widely especially for disabled people. Vibrotactile stimuli are alternatively used in BCI especially for patients whose vision or eye movements are impaired. Moreover, users' training in tactile BCI is also easy. In information theory and coding theory, error-detecting codes are able to detect errors from the received data which is transmitted over unreliable transmission channels. BCIs could use the advantage of error-detecting codes because the classification process could be considered as a noisy transmission channel from translating brain waves to the user's intent. In this paper, we present a P300 vibrotactile BCI based on the electroencephalogram (EEG) with error-detecting codes. A parity check code which is a simple method of error detection is used as an error-detecting code. The aim of this study is to compare the efficiency between a vibrotactile BCI with and without error-detecting codes. The classification accuracy and the information transfer rate (ITR) of applying error detection are improved 12.04 % and 0.53 bit/min, respectively, compared to the conventional method which does not apply error-correcting codes.
研究論文（国際会議プロシーディングス）, 英語 - Neural Decoding of Code Modulated Visual Evoked Potentials by Spatio-Temporal Inverse Filtering for Brain Computer Interfaces
Jun-ichi Sato; Yoshikazu Washizawa
2016 38TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC), IEEE, 1巻, 1号, 掲載ページ 1484-1487, 出版日 2016年, 査読付, This study addresses neural decoding of a code modulated visual evoked potentials (c-VEPs). c-VEP was recently developed, and applied to brain computer interfaces (BCIs). c-VEP BCI exhibits faster communication speed than existing VEP-based BCIs. In c-VEP BCI, the canonical correlation analysis (CCA) that maximizes the correlation between an averaged signal and single trial signals is often used for the spatial filter. However, CCA does not utilize information of given PN sequence, and hence, the filtered signal may not have properties of PN sequence. In this paper, we propose a decoding method to restore the given PN sequence from the observed VEP. We compare linear and non-linear spatio-temporal inverse filtering methods. For the linear method, the least mean square error and lasso are used to obtain the filter coefficients. For the non-linear method, the artificial neural network is used. The proposed methods exhibited better decoding performance, and higher classification accuracies than conventional CCA spatial filtered c-VEP BCI.
研究論文（国際会議プロシーディングス）, 英語 - Reliability-based Automatic Repeat Request for Short Code Modulation Visual Evoked Potentials in Brain Computer Interfaces
Jun-ichi Sato; Yoshikazu Washizawa
2015 37TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC), IEEE, 掲載ページ 562-565, 出版日 2015年, 査読付, We propose two methods to improve code modulation visual evoked potential brain computer interfaces (cVEP BCIs). Most of BCIs average brain signals from several trials in order to improve the classification performance. The number of averaging defines the trade-off between input speed and accuracy, and the optimal averaging number depends on individual, signal acquisition system, and so forth. Firstly, we propose a novel dynamic method to estimate the averaging number for cVEP BCIs. The proposed method is based on the automatic repeat request (ARQ) that is used in communication systems. The existing cVEP BCIs employ rather longer code, such as 63-bit M-sequence. The code length also defines the trade-off between input speed and accuracy. Since the reliability of the proposed BCI can be controlled by the proposed ARQ method, we introduce shorter codes, 32-bit M-sequence and the Kasami-sequence. Thanks to combine the dynamic averaging number estimation method and the shorter codes, the proposed system exhibited higher information transfer rate compared to existing cVEP BCIs.
研究論文（国際会議プロシーディングス）, 英語 - METRICS OF GRASSMANNIAN REPRESENTATION IN REPRODUCING KERNEL HILBERT SPACE FOR VARIATIONAL PATTERN ANALYSIS
Yoshikazu Washizawa
2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP), IEEE, 掲載ページ 2194-2198, 出版日 2015年, 査読付, Variation of patterns in signal can be represented by the covariance structure of vectors or its eigensubspace. When information of the pattern variation is available, representation by the covariance matrix or the eigensubspace is useful for feature extraction and classification compared with standard vector or matrix representations.
The structure and metric of the Grassmann manifold (Grassmannian) which is a set of eigensubspace, have been researched widely. Especially, the author has developed Mahalanobis distance in the Grassmannian, and it shows higher representation ability and classification accuracy compared with conventional Grassmannian representation methods.
In this paper, we extend Grassmannian metrics including the Mahalanobis distance using the kernel trick. We also propose an efficient basis vector selection algorithm and combine with the subset approximation of kernel principal component analysis to reduce the computational cost. In our experimental simulation by 3-dimensional object recognition problem, the proposed Mahalanobis distance shows better performance than conventional methods.
研究論文（国際会議プロシーディングス）, 英語 - Spatial filter for short period code modulation visual evoked potentials in brain computer interfaces
J. Sato; Y. Washizawa
Proceedings of the 30th International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC 2015), 掲載ページ 23-26, 出版日 2015年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Sign Language Recognition with Microsoft Kinect's Depth and Colour Sensors
Panupon Usachokcharoen; Yoshikazu Washizawa; Kitsuchart Pasupa
2015 IEEE INTERNATIONAL CONFERENCE ON SIGNAL AND IMAGE PROCESSING APPLICATIONS (ICSIPA), IEEE, accepted巻, 掲載ページ 186-190, 出版日 2015年, 査読付, In the last few years, many technologies for helping differently-abled people have been developed continually including technologies for recognising sign language that enables them to communicate with each other. In this research, we studied sign language recognition using Microsoft Kinect. Conventionally, Microsoft Kinect uses its depth sensor to collect depth and motion features in order to recognise words in sign language. Our proposed method improved it by adding colour feature sensing. Acquired by the depth and colour sensors, all of the features were extracted and then machine-learned by multi-class Support Vector Machine. The learned features were associated with the following words: 'Name', 'No', 'Thank you', 'How many', 'What', 'Where', 'Yes', and 'Your'. An experiment to find out which combination of the three features-depth, motion, and colour-predicted the mentioned words most accurately showed that the combination of motion and colour features achieved the highest accuracy at 95%.
研究論文（国際会議プロシーディングス）, 英語 - Channel Selection for Brain Signal Classification by Penalized Automatic Relevance Determination
Reo Togashi; Yoshikazu Washizawa
2015 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA), IEEE, accepted巻, 掲載ページ 1023-1027, 出版日 2015年, 査読付, Channel selection or reduction in Brain computer interface (BCI) is important to reduce the cost and improve the generalized accuracy. A channel selection method using group automatic relevance determination (GARD) for P300 based BCI has been reported. In this paper, we apply the penalized ARD (PARD) which is an extension of ARD, and compare with GARD in our auditory BCI. Experimental results show that PARD provides more sparse solution than GARD while PARD shows almost the same classification accuracy as GARD.
研究論文（国際会議プロシーディングス）, 英語 - Mahalanobis Distance on Extended Grassmann Manifolds for Variational Pattern Analysis
Yoshikazu Washizawa; Seiji Hotta
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 25巻, 11号, 掲載ページ 1980-1990, 出版日 2014年11月, 査読付, In pattern classification problems, pattern variations are often modeled as a linear manifold or a low-dimensional subspace. Conventional methods use such models and define a measure of similarity or dissimilarity. However, these similarity measures are deterministic and do not take into account the distribution of linear manifolds or low-dimensional subspaces. Therefore, if the distribution is not isotopic, the distance measurements are not reliable, as well as vector-based distance measurement in the Euclidean space. We previously systematized the representations of variational patterns using the Grassmann manifold and introduce the Mahalanobis distance to the Grassmann manifold as a natural extension of Euclidean case. In this paper, we present two methods that flexibly extend the Mahalanobis distance on the extended Grassmann manifolds. These methods can be used to measure pattern (dis)similarity on the basis of the pattern structure. Experimental evaluation of the performance of the proposed methods demonstrated that they exhibit a lower error classification rate.
研究論文（学術雑誌）, 英語 - Multiple kernel learning for quadratically constrained MAP classification
Yoshikazu Washizawa; Tatsuya Yokota; Yukihiko Yamashita
IEICE Transactions on Information and Systems, Institute of Electronics, Information and Communication, Engineers, IEICE, E96-D巻, 5号, 掲載ページ 1340-1344, 出版日 2014年, 査読付, Most of the recent classification methods require tuning of the hyper-parameters, such as the kernel function parameter and the regularization parameter. Cross-validation or the leave-one-out method is often used for the tuning, however their computational costs aremuch higher than that of obtaining a classifier. Quadratically constrained maximum a posteriori (QCMAP) classifiers, which are based on the Bayes classification rule, do not have the regularization parameter, and exhibit higher classification accuracy than support vector machine (SVM). In this paper, we propose a multiple kernel learning (MKL) for QCMAP to tune the kernel parameter automatically and improve the classification performance. By introducing MKL, QCMAP has no parameter to be tuned. Experiments show that the proposed classifier has comparable or higher classification performance than conventional MKL classifiers. © 2014 The Institute of Electronics, Information and Communication Engineers.
研究論文（学術雑誌）, 英語 - A novel EEG-based spelling system using N100 and P300
Hikaru Sato; Yoshikazu Washizawa
Studies in Health Technology and Informatics, IOS Press, 205巻, 掲載ページ 428-432, 出版日 2014年, 査読付, P300-speller is one of the most popular EEG-based spelling systems proposed by Farwell and Donchin. P300-speller has a 6×6 character matrix and requires at least 12 flashes to input one character. This restricts increasing of the information transfer rate (ITR) by decreasing the number of flashes. In this paper, a novel spelling system is proposed to reduce the number of flashes by sequential stimulation of images. In order to determine the command, the proposed system utilizes two kinds of the event-related brain potentials (ERP), N100 and P300 whereas P300-speller utilizes only P300. Thanks to using both N100 and P300, the proposed system achieves higher accuracy and faster spelling speed than P300-speller. Our experiment by ten subjects showed that ITR of the proposed system is an average of 0.25bit/sec improved compared to P300-speller.
研究論文（国際会議プロシーディングス）, 英語 - N100-P300 Speller BCI with detection of user's input intention
H. Sato; Y. Washizawa
Proc. of 6th International Brain-Computer Interface Conference, 掲載ページ 242-245, 出版日 2014年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Bayesian delay time estimation of brain signal using N100 response for auditory BCI
Reo Togashi; Yoshikazu Washizawa
2014 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA), IEEE, accepted巻, 出版日 2014年, 査読付, Brain computer interface (BCI) enables disabled people to communicate by brain signal. P300 which appears 300ms after the onset of a low frequent stimulus is extensively used to actualize BCI. Precise detection of P300 component is therefore important. Most of existing BCI assumes that P300 is observed after 300ms, however this latency has variation due to the condition of a subject and the level of attention for the stimulus. This latency variation distorts averaged P300 and hence incurs the deterioration of the classification accuracy. A delay time estimation method for P300 signal using Bayesian estimation has been reported in the previous study to address this problem. However, the method has a problem that the algorithm fails to estimate the delay time when the signal does not contain P300. A Bayesian delay time estimation method using N100 component is therefore proposed. This proposed method exhibited 3.2% higher classification accuracy than the conventional delay time estimation method in auditory BCI.
研究論文（国際会議プロシーディングス）, 英語 - Bayesian delay time estimation of brainwaves using N100 response in tactile-force brain-computer interface
R. Togashi; Y. Washizawa
Proc of SCIS and ISIS, accepted巻, 掲載ページ 1-1, 出版日 2014年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Electroencephalographic gamma-band activity and music perception in musicians and non-musicians
Y. Urakami; K. Kawamura; Y. Washizawa; A. Cichocki
Activitas Nervosa Superior Rediviva, 44巻, 4号, 掲載ページ 149-157, 出版日 2013年12月, 査読付
研究論文（学術雑誌）, 英語 - 音楽認知におけるγ活動の意義―意識・認知との関連から
浦上裕子; 川村光毅; 鷲沢嘉一; 日吉和子; アンジェイ チホツキ
臨床神経生理学会誌, Japanese Society of Clinical Neurophysiology, 41巻, 3号, 掲載ページ 209-219, 出版日 2013年08月, 査読付, 音を聴覚的に認識し, ハーモニー, 音韻, 旋律など音楽を認知することで前頭前野が活動し行動が起こる。本研究は音楽大学学生5名,音楽専門家１名, 非音楽大学生5名 (21–25歳: 男9女2) を対象とし60 ch脳波を用いて安静閉眼時, 「新世界より」「レクイエム」聴取時, イメージ時の脳活動を計測し, 音楽認知の神経基盤を明らかにすることを目的とした。Morlet waveletによる時間周波数解析を行い, 各周波数帯域の平均信号強度をroot mean squareとして求め成分比較を行った。音楽聴取時は無音安静閉眼時に比べ, δ, α, β, γ活動に有意な減少を認めた。γ活動の減少が最も大きく, 全脳部位で有意に減少した。音楽家は音楽聴取時・イメージ時ともに前頭部γ活動が減少, 非音楽家はイメージ時には前頭部γ活動が増加した。音楽家と非音楽家のγ活動の差は, 音楽という経験による意識や注意, 情動の統合や潜在記憶の差を反映する可能性がある。
研究論文（学術雑誌）, 日本語 - Feature Extraction of P300 Signal Using Bayesian Delay Time Estimation
Reo Togashi; Yoshikazu Washizawa
2013 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA), IEEE, 233号, 掲載ページ 1-5, 出版日 2013年, 査読付, Brain-computer interfaces (BCIs) based on event-related potentials (ERP) are communicating tools with severely disabled people. P300 which is observed after 300 mili seconds from stimuli is widely used for the operation principle of BCIs. However the response time to the stimuli depends on a subject, trial, and also a channel. Many existing approaches ignore this variation and extract only low frequency component. We propose a method to estimate the response time of P300 using Bayesian estimation. The proposed method exhibited higher performance in our auditory BCI.
研究論文（国際会議プロシーディングス）, 英語 - Gamma-band activity; role of music perception in the brain
Yuko Urakami; Yoshikazu Washizawa; Koki Kawamura; Kazuko Hiyoshi; Andrzej Cichocki
Proc. of the 19th Annual Meeting of the Organization for Human Brain Mapping, 3557号, 出版日 2013年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Single trial BCI classification accuracy improvement for the novel virtual sound movement-based spatial auditory paradigm
Yohann Lelievre; Yoshikazu Washizawa; Tomasz M. Rutkowski
2013 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA 2013, 365号, 出版日 2013年, 査読付, This paper presents a successful attempt to improve single trial P300 response classification results in a novel moving sound spatial auditory BCI paradigm. We present a novel paradigm, together with a linear support vector machine classifier application, which allows a boost in single trial based spelling accuracy in comparison with classic stepwise linear discriminant analysis methods. The results of the offline classification of the P300 responses of seven subjects support the proposed concept, with a classification improvement of up to 80%, leading, in the best case presented, to an information transfer rate boost of 28.8 bit/min. © 2013 APSIPA.
研究論文（国際会議プロシーディングス）, 英語 - Gamma-activity in the frontal lobe; represent self-consciousness and sustained attention of musicians and traumatic brain injured patients,
Y. Urakami; K. Kawamura; Y. Washizawa; K. Hiyoshi; A. Cichocki
Joint Annual Meeting ECNS-ISBET-ISNIP-EPIC, 掲載ページ 36, 出版日 2013年
研究論文（国際会議プロシーディングス）, 英語 - On Investigating Efficient Methodology for Environmental Sound Recognition
Cruz Alfredo Ruiz-Martinez; Muhammad Tahir Akhtar; Yoshikazu Washizawa; Enrique Escamilla-Hernandez
2013 INTERNATIONAL SYMPOSIUM ON INTELLIGENT SIGNAL PROCESSING AND COMMUNICATIONS SYSTEMS (ISPACS), IEEE, 掲載ページ 210-214, 出版日 2013年, 査読付, This paper presents a comparative study of various methods to identify the environmental sounds. We evaluate two methods for feature extraction: Mel Frequency Cepstral Coefficients (MFCC) which is well known for speaker identification, and Matching Pursuit (MP) with Gabor Dictionary which gives a time frequency representation employed for scene recognition. In the classification stage, we show a comparison among Support Vector Machines (SVM), Logistic Regression, and Backpropagation Artificial Neural Network (BP-ANN). Simulation results show that MFCC gives a higher recognition performance as compared with MP. Furthermore, by concatenating MFCC features with some feature of MP, e.g., scale, might also improve performance in some situations. We observe that SVM show the best performance among the classifiers, for clean as well noisy signals.
研究論文（国際会議プロシーディングス）, 英語 - Gamma-band Synchrony; Role of Music Perception in the Brain.
Urakami Y; Washizawa Y; Kawamura K; Hiyoshi K; Cichocki A
International conference of Electroencephalography and Clinical Neuroscience Society, 掲載ページ 11-15, 出版日 2012年12月, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Adaptive Subset Kernel Principal Component Analysis for Time-Varying Patterns
Yoshikazu Washizawa
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 23巻, 12号, 掲載ページ 1961-1973, 出版日 2012年12月, 査読付, Kernel principal component analysis (KPCA) and its online learning algorithms have been proposed and widely used. Since KPCA uses training samples for bases of the operator, its online learning algorithms require the preparation of all training samples beforehand. Subset KPCA (SubKPCA), which uses a subset of samples for the basis set, has been proposed and has demonstrated better performance with less computational complexity. In this paper, we extend SubKPCA to an online version and propose methods to add and exchange a sample in the basis set. Since the proposed method uses the basis set, we do not need to prepare all training samples beforehand. Therefore, the proposed method can be applied to time-varying patterns, in contrast to existing online KPCA algorithms. Experimental results demonstrate the advantages of the proposed method.
研究論文（学術雑誌）, 英語 - ADAPTIVE KERNEL PRINCIPAL COMPONENTS TRACKING
Toshihisa Tanaka; Yoshikazu Washizawa; Anthony Kuh
2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), IEEE, 掲載ページ 1905-1908, 出版日 2012年, 査読付, Adaptive online algorithms for simultaneously extracting nonlinear eigenvectors of kernel principal component analysis (KPCA) are developed. KPCA needs all the observed samples to represent basis functions, and the same scale of eigenvalue problem as the number of samples should be solved. This paper reformulates KPCA and deduces an expression in the Euclidean space, where an algorithm for tracking generalized eigenvectors is applicable. The developed algorithm here is least mean squares (LMS)-type and recursive least squares (RLS)-type. Numerical example is then illustrated to support the analysis.
研究論文（国際会議プロシーディングス）, 英語 - MAHALANOBIS DISTANCE ON GRASSMANN MANIFOLD AND ITS APPLICATION TO BRAIN SIGNAL PROCESSING
Yoshikazu Washizawa; Seiji Hotta
2012 IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING (MLSP), IEEE, 0巻, 0号, 掲載ページ 0, 出版日 2012年, 査読付, Multi-dimensional data such as image patterns, image sequences, and brain signals, are often given in the form of the variance-covariance matrices or their eigenspaces to represent their own variations. For example, in face or object recognition problems, variations due to illuminations, camera angles can be represented by eigenspaces. A set of the eigenspaces is called the Grassmann manifold, and simple distance measurements in the Grassmann manifold, such as the projection metric have been used in conventional researches. However, in linear spaces, if the distribution of patterns is not isotropic, statistical distances such as the Mahalanobis distance are reasonable, and their performances are higher than simple distances in many problems. In this paper, we introduce the Mahalanobis distance in the Grassmann manifolds. Two experimental results, an object recognition problem and a brain signal processing, demonstrate the advantages of the proposed distance measurement.
研究論文（国際会議プロシーディングス）, 英語 - Toward multi-command auditory brain computer interfacing using speech stimuli
Shuho Yoshimoto; Yoshikazu Washizawa; Toshihisa Tanaka; Hiroshi Higashi; Jun Tamura
2012 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA ASC), IEEE, 掲載ページ 1-4, 出版日 2012年, 査読付, Brain-computer interfaces (BCIs) based on event-related potentials (ERP) are promising tools to communicate with patients suffering from some severe disabled diseases. ERP is evoked by various stimuli such as auditory, olfactory, and visual stimuli. Some auditory based BCIs with certain synthetic tone have been proposed, however, it is still challenging to increase the number of commands in auditory-based BCIs, since it is usually difficult for users to remember and distinguish multiple tones that corresponds to commands. We propose a new auditory BCI framework using speech stimuli. It is easier for users to distinguish different speech stimuli than different simple tones. We show experimental results of four-command BCI. The proposed speech-based BCI achieved a classification accuracy of more than 70 percents.
研究論文（国際会議プロシーディングス）, 英語 - Imagery Movement Paradigm User Adaptation Improvement with Quasi-movements Phenomenon
Hiroshi Higashi; Tomasz M. Rutkowski; Yoshikazu Washizawa; Toshihisa Tanaka; Andrzej Cichocki
ADVANCES IN COGNITIVE NEURODYNAMICS (II), SPRINGER-VERLAG BERLIN, 2巻, 掲載ページ 677-681, 出版日 2011年, We discuss a novel movement imagery brain-computer/ma-chine-interface (BCl/BMI) paradigm learning procedure with utilization of real- and quasi-movements of subjects' thumbs. In the proposed procedure volitional movements are slowly minimized by the subjects to a very low level so that finally they become undetectable by objective measures such as electromyography (EMG). The procedure allows the subjects to understand motion imagery process, which follows after the training. The procedure allows also to control the final movement imagery protocol and to detect any possible movements in case subject would not learn to suppress them completely. We present also a discussion on electroencephalography (EEG) signals pre-processing steps with common spatial pattern (CSP) method improvements. Promising results were obtained with subjects who could not perform the motion imagery paradigm as well with those who never tried it before conclude the paper.
研究論文（国際会議プロシーディングス）, 英語 - Centered Subset Kernel PCA for Denoising
Yoshikazu Washizawa; Masayuki Tanaka
COMPUTER VISION - ACCV 2010 WORKSHOPS, PT II, SPRINGER-VERLAG BERLIN, 6469巻, 掲載ページ 354-363, 出版日 2011年, 査読付, Kernel PCA has been applied to image processing, even though, it is known to have high computational complexity. We introduce centered Subset KPCA for image denoising problems. Subset KPCA has been proposed for reduction of computational complexity of KPCA, however, it does not consider a pre-centering that is often important for image processing. Indeed, pre-centering of Subset KPCA is not straightforward because Subset KPCA utilizes two sets of samples. We propose an efficient algorithm for pre-centering, and provide an algorithm for pre-image. Experimental results show that our method is comparable with a state-of-the-art image denoising method.
研究論文（国際会議プロシーディングス）, 英語 - Trace Norm Regularization and Application to Tensor Based Feature Extraction
Yoshikazu Washizawa
COMPUTER VISION - ACCV 2010 WORKSHOPS, PT II, SPRINGER-VERLAG BERLIN, 6469巻, 掲載ページ 404-413, 出版日 2011年, 査読付, The trace norm regularization has an interesting property that is rank of a matrix is reduced according to its continuous regularization parameter. We propose a new efficient algorithm for a kind of trace norm regularization problems. Since the algorithm is not gradient-based approach, its computational complexity does not depend on initial states or learning rate. We also apply the proposed algorithm to a tensor based feature extraction method, that is an extension of the trace norm regularized feature extraction.
Computational simulations show that the proposed algorithm provides an accurate solution in less time than conventional methods. The proposed trace based feature extraction method show almost that same performance as Multilinear PCA.
研究論文（国際会議プロシーディングス）, 英語 - EEG auditory steady state responses classification for the novel BCI
Hiroshi Higashi; Tomasz M. Rutkowski; Yoshikazu Washizawa; Andrzej Cichocki; Toshihisa Tanaka
Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS, 掲載ページ 4576-4579, 出版日 2011年, 査読付, An auditory modality brain computer interface (BCI) is a novel and interesting paradigm in neurotechnology applications. The paper presents a concept of auditory steady state responses (ASSR) utilization for the novel BCI paradigm. Two EEG feature extraction approaches based on a bandpass filtering and an AR spectrum estimation are tested together with two classification schemes in order to validate the proposed auditory BCI paradigm. The resulting good classification scores of users intentional choices, of attending or not to the presented stimuli, support the hypothesis of the ASSR stimuli validity for a solid BCI paradigm. © 2011 IEEE.
研究論文（国際会議プロシーディングス）, 英語 - Kernel Wiener Filter and its Application to Pattern Recognition
Hirokazu Yoshino; Chen Dong; Yoshikazu Washizawa; Yukihiko Yamashita
IEEE TRANSACTIONS ON NEURAL NETWORKS, IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 21巻, 11号, 掲載ページ 1719-1730, 出版日 2010年11月, 査読付, The Wiener filter (WF) is widely used for inverse problems. From an observed signal, it provides the best estimated signal with respect to the squared error averaged over the original and the observed signals among linear operators. The kernel WF (KWF), extended directly from WF, has a problem that an additive noise has to be handled by samples. Since the computational complexity of kernel methods depends on the number of samples, a huge computational cost is necessary for the case. By using the first-order approximation of kernel functions, we realize KWF that can handle such a noise not by samples but as a random variable. We also propose the error estimation method for kernel filters by using the approximations. In order to show the advantages of the proposed methods, we conducted the experiments to denoise images and estimate errors. We also apply KWF to classification since KWF can provide an approximated result of the maximum a posteriori classifier that provides the best recognition accuracy. The noise term in the criterion can be used for the classification in the presence of noise or a new regularization to suppress changes in the input space, whereas the ordinary regularization for the kernel method suppresses changes in the feature space. In order to show the advantages of the proposed methods, we conducted experiments of binary and multiclass classifications and classification in the presence of noise.
研究論文（学術雑誌）, 英語 - Blind Extraction of Global Signal From Multi-Channel Noisy Observations
Yoshikazu Washizawa; Yukihiko Yamashita; Toshihisa Tanaka; Andrzej Cichocki
IEEE TRANSACTIONS ON NEURAL NETWORKS, IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 21巻, 9号, 掲載ページ 1472-1481, 出版日 2010年09月, 査読付, We propose a novel efficient method of blind signal extraction from multi-sensor networks when each observed signal consists of one global signal and local uncorrelated signals. Most of existing blind signal separation and extraction methods such as independent component analysis have constraints such as statistical independence, non-Gaussianity, and underdetermination, and they are not suitable for global signal extraction problem from noisy observations. We developed an estimation algorithm based on alternating iteration and the smart weighted averaging. The proposed method does not have strong assumptions such as independence or non-Gaussianity. Experimental results using a musical signal and a real electroencephalogram demonstrate the advantage of the proposed method.
研究論文（学術雑誌）, 英語 - Feature Extraction Using Constrained Approximation and Suppression
Yoshikazu Washizawa
IEEE TRANSACTIONS ON NEURAL NETWORKS, IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 21巻, 2号, 掲載ページ 201-210, 出版日 2010年02月, 査読付, In this paper, we systematize a family of constrained quadratic classifiers that belong to the class of one-class classifiers. One-class classifiers such as the single-class support vector machine or the subspace methods are widely used for pattern classification and detection problems because they have many advantages over binary classifiers. We interpret subspace methods as rank-constrained quadratic classifiers in the framework. We also introduce two constraints and a method of suppressing the effect of competing classes to make them more accurate and retain their advantages over binary classifiers. Experimental results demonstrate the advantages of our methods over conventional classifiers.
研究論文（学術雑誌）, 英語 - Tensor based simultaneous feature extraction and sample weighting for EEG classification
Yoshikazu Washizawa; Hiroshi Higashi; Tomasz Rutkowski; Toshihisa Tanaka; Andrzej Cichocki
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 6444 LNCS巻, PART 2号, 掲載ページ 26-33, 出版日 2010年, 査読付, In this paper we propose a Multi-linear Principal Component Analysis (MPCA) which is a new feature extraction and sample weighting method for classification of EEG signals using tensor decomposition. The method has been successfully applied to Motor-Imagery Brain Computer Interface (MI-BCI) paradigm. The performance of the proposed approach has been compared with standard Common Spatial Pattern (CSP) as well with a combination of PCA and CSP methods. We have achieved an average accuracy improvement of two classes classification in a range from 4 to 14 percents. © 2010 Springer-Verlag.
研究論文（国際会議プロシーディングス）, 英語 - Subset kernel PCA for pattern recognition
Yoshikazu Washizawa
2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops 2009, 掲載ページ 162-169, 出版日 2009年, 査読付, Subspace methods that utilize principal component analysis (PCA) are widely used for pattern classification or detection problems. Kernel PCA (KPCA) that is an extension of PCA is also applied to subspace methods. However, its computational cost is very high since the computational cost mainly depends on the number of samples in kernel methods. Recently, subset KPCA (SKPCA) has been proposed in order to reduce its computational complexity. In this paper, we apply SKPCA to subspace methods, and compare SKPCA with KPCA using some sample selection methods. Experimental results demonstrate advantages of subspace methods using SKPCA. ©2009 IEEE.
研究論文（国際会議プロシーディングス）, 英語 - SUBSET KERNEL PRINCIPAL COMPONENT ANALYSIS
Yoshikazu Washizawa
2009 IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING, IEEE, 掲載ページ 357-362, 出版日 2009年, 査読付, Kernel principal component analysis (kernel PCA or KPCA) has been used widely for non-linear feature extraction, dimensionally reduction, and classification problems. However, KPCA is known to have high computational complexity, that is the eigenvalue decomposition of which size equals to the number of samples n. Moreover, in order to calculate projection of vector onto the subspace obtained by KPCA, we have to store all n samples and evaluate the kernel function n times. In order to overcome these problems, we propose subset KPCA that minimizes a residual error for all samples using limited number of them, and we provide its solution. Experimental results using synthetic and real data show that the proposed method gives almost the same result as KPCA even if the size of the problem is one-tenth of KPCA.
研究論文（国際会議プロシーディングス）, 英語 - Pattern classification on local metric structure
Yoshikazu Washizawa
Proceedings of the International Conference on Document Analysis and Recognition, ICDAR, 掲載ページ 471-475, 出版日 2009年, 査読付, A metric is an important concept in pattern classification problems. Many metrics have been applied to pattern classification problems, e.g., the Mahalanobis distance or shift-invariant distance. However, a metric is not uniform in whole domain, in other words, structure of patterns are different in each local domain. Several approaches that utilize such local structure have been proposed. In this paper, we systematize them and propose a framework to describe patterns by a d-dimensional vector and local metric matrix at the point. Then, we introduce two distance measurements to this framework. Experimental results demonstrate advantages of the proposed methods. © 2009 IEEE.
研究論文（国際会議プロシーディングス）, 英語 - Blind source extraction using spatio-temporal inverse filter
Yoshikazu Washizawa; Yukihiko Yamashita; Andrzej Cichocki
ISCAS: 2009 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-5, IEEE, 掲載ページ 2786-+, 出版日 2009年, 査読付, Blind source extraction is one of the most important problems for multi-sensor networks. We propose a blind source extraction and deconvolution method in the presence of noise. We use MA-model for the signal generation model, and the convolutive observation model. The parameter of MA-model and the observations are obtained from an alternating least square (ALS) algorithm. The reconstruction is done by an spatio-temporal inverse filter such that it minimizes the Euclidean distance between the original signal and the reconstruction signal. Experimental results demonstrate advantages of the proposed method.
研究論文（国際会議プロシーディングス）, 英語 - Noninvasive BCIs: Multiway signal-processing array decompositions
Andrzej Cichocki; Yoshikazu Washizawa; Tomasz Rutkowski; Hovagim Bakardjian; Anh Huy Phan; Seungjin Choi; Hyekyoung Lee; Qibin Zhao; Liqing Zhang; Yuanqing Li
Computer, 41巻, 10号, 掲載ページ 34-42, 出版日 2008年, 査読付, In addition to helping better understand how the human brain works, the brain-computer interface neuroscience paradigm allows researchers to develop a new class of bioengineering control devices and robots, offering promise for rehabilitation and other medical applications as well as exploring possibilities for advanced human-computer interfaces. © 2008 IEEE.
研究論文（学術雑誌）, 英語 - Robust boundary learning for multi-class classification problems
Y. Washizawa; S. Hotta
in Proc. of 1st IAPR Workshop on Cognitive Information Processing (CIP 2008), 掲載ページ 188-193, 出版日 2008年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Blind global source extraction from noisy observations
Y. Washizawa; Y. Yamashita; T. Tanaka; A. Cichocki
in Proc. of 2008 RISP International Workshop on Nonlinear Circuits and Signal Processing (NCSP'08), 掲載ページ 184-187, 出版日 2008年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Simultaneous pattern classification and multidomain association using self-structuring kernel memory networks
T. Hoya; Y. Washizawa
IEEE Trans. Neural Networks, 18巻, 3号, 掲載ページ 732-744, 出版日 2007年03月, 査読付
研究論文（学術雑誌）, 英語 - Extraction of steady state visually evoked potential signal and estimation of distribution map from EEG data.
Washizawa Y; Yamashita Y; Tanaka T; Cichocki A
Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, 2007巻, 掲載ページ 5449-5452, 出版日 2007年, 査読付 - Sparse blind identification and separation by using adaptive K-orthodrome clustering
Y. Washizawa; A. Cichocki
Neurocomputing, 71巻, 10-12号, 掲載ページ 2321-2329, 出版日 2007年, 査読付
研究論文（学術雑誌）, 英語 - Extraction of steady state visually evoked potential signal and estimation of distribution map from EEG data
Yoshikazu Washizawa; Yukihiko Yamashita; Tohihisa Tanaka; Andrzej Cichocki
2007 ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY, VOLS 1-16, IEEE, 掲載ページ 5449-5452, 出版日 2007年, 査読付, We propose a signal extraction method from multi-channel EEG signals and apply to extract Steady State Visually Evoked Potential (SSVEP) signal. SSVEP is a response to visual stimuli presented in the form of flushing patterns. By using several flushing patterns with different frequency, brain machine (computer) interface (BMI/BCI) can be realized. Therefore it is important to extract SSVEP signals from multi-channel EEG signals.
At first, we estimate the power of the objective signal in each electrode. Estimation of the power is helpful in not only extraction of the signal but also drawing a distribution map of the signal, finding electrodes which have large SNR, and ranking electrodes in sort of information with respect to the power of the signal.
Experimental results show that the proposed method 1) estimates more accurate power than existing methods, 2) estimates the global signal which has larger SNR than existing methods, and 3) allows us to draw a distribution map of the signal, and it conforms the biological theory.
研究論文（国際会議プロシーディングス）, 英語 - Regularization vs. rank reduction in quadratic classifiers
Y. Washizawa
Proc. of 8th Asian Conference on Computer Vision (ACCV 2007), 掲載ページ 108-115, 出版日 2007年, 査読付
研究論文（国際会議プロシーディングス）, 英語 - Kernel projection classifiers with suppressing features of other classes
Y Washizawa; Y Yamashita
NEURAL COMPUTATION, M I T PRESS, 18巻, 8号, 掲載ページ 1932-1950, 出版日 2006年08月, 査読付, We propose a new classification method based on a kernel technique called suppressed kernel sample space projection classifier (SKSP), which is extended from kernel sample space projection classifier (KSP).
In kernel methods, samples are classified after they are mapped from an input space to a high-dimensional space called a feature space. The space that is spanned by samples of a class in a feature space is defined as a kernel sample space. In KSP, an unknown input vector is classified to the class of which projection norm onto the kernel sample space is maximized. KSP can be interpreted as a special type of kernel principal component analysis (KPCA). KPCA is also used in classification problems. However, KSP has more useful properties compared with KPCA, and its accuracy is almost the same as or better than that of KPCA classifier.
Since KSP is a single-class classifier, it uses only self-class samples for learning. Thus, for a multiclass classification problem, even though there are very many classes, the computational cost does not change for each class. However, we expect that more useful features can be obtained for classification if samples from other classes are used. By extending KSP to SKSP, the effects of other classes are suppressed, and useful features can be extracted with an oblique projection.
Experiments on two-class classification problems, indicate that SKSP shows high accuracy in many classification problems.
研究論文（学術雑誌）, 英語 - On-line K-plane clustering learning algorithm for sparse comopnent analysis
Yoshikazu Washizawa; Andrzej Cichocki
2006 IEEE International Conference on Acoustics, Speech and Signal Processing, Vols 1-13, IEEE, 掲載ページ 5539-5542, 出版日 2006年, 査読付, In this paper we propose a new algorithm for identifying mixing (basis) matrix A knowing only sensor (data) matrix X for linear model X = AS + E, under some weak or relaxed conditions, expressed in terms of sparsity of latent (hidden) components represented by the matrix S. We present a simple and efficient on-line algorithm for such identification and illustrate its performance by estimation of unknown matrix A and source signals S. The main feature of the proposed algorithm is its adaptivity to changing environment and robustness in respect to noise and outliers that do not satisfy sparseness conditions.
研究論文（国際会議プロシーディングス）, 英語 - A flexible method for envelope estimation in empirical mode decomposition
Yoshikazu Washizawa; Toshihisa Tanaka; Danilo P. Mandic; Andrzej Cichocki
KNOWLEDGE-BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 3, PROCEEDINGS, SPRINGER-VERLAG BERLIN, 4253巻, 掲載ページ 1248-1255, 出版日 2006年, 査読付, A flexible and efficient method for finding the envelope within the empirical mode decomposition (EMD) is introduced. Unlike the existing (deterministic) spline based strategy, the proposed envelope is a result of an optimisation precess and sought as a minimum of a quadratic cost function. A closed form solution of this optimisation problem is obtained and it is shown that by choosing free parameters, we can fine-tune the frequency resolution or the number of intrinsic mode functions (IMFs) as well as the shape of the envelopes. Computer simulations on both the synthetic and real-world electro-encephalogram (EEG) data support the analysis.
研究論文（学術雑誌）, 英語 - Non-linear Wiener filter in reproducing kernel Hilbert space
Yoshikazu Washizawa; Yukihiko Yamashita
18TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOL 1, PROCEEDINGS, IEEE COMPUTER SOC, 掲載ページ 967-+, 出版日 2006年, 査読付, Wiener filters are used widely for inverse problems. From an observed signal, a Wiener filter provides the best restored signal with respect to the square error averaged over the original signal and the noise among linear operators. We introduce the non-linear Wiener filter, which is a kernel-based extension of the Wiener filter When the kernel method is applied to the Wiener filter directly, the dimensions of the space where the calculation has to be done is very large since noise samples have to be used. We provide a realistic solution using the first order approximation. Moreover, we provide the experimental results to demonstrate the advantages of this method.
研究論文（国際会議プロシーディングス）, 英語 - On-line K-plane clustering learning algorithm for sparse comopnent analysis
Yoshikazu Washizawa; Andrzej Cichocki
2006 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL V, PROCEEDINGS, IEEE, 掲載ページ 681-+, 出版日 2006年, 査読付, In this paper we propose a new algorithm for identifying mixing (basis) matrix A knowing only sensor (data) matrix X for linear model X = AS + E, under some weak or relaxed conditions, expressed in terms of sparsity of latent (hidden) components represented by the matrix S. We present a simple and efficient on-line algorithm for such identification and illustrate its performance by estimation of unknown matrix A and source signals S. The main feature of the proposed algorithm is its adaptivity to changing environment and robustness in respect to noise and outliers that do not satisfy sparseness conditions.
研究論文（国際会議プロシーディングス）, 英語 - Kernel relative principal component analysis for pattern recognition
Y Washizawa; K Hikida; T Tanaka; Y Yamashita
STRUCTURAL, SYNTACTIC, AND STATISTICAL PATTERN RECOGNITION, PROCEEDINGS, SPRINGER-VERLAG BERLIN, 3138巻, 掲載ページ 1105-1113, 出版日 2004年, 査読付, Principal component analysis (PCA) is widely used in signal processing, pattern recognition, etc. PCA was extended to the relative PCA (RPCA). RPCA provides principal components of a signal while suppressing effects of other signals. PCA was also extended to the kernel PCA (KPCA). By using a mapping from the original space to a higher dimensional space and its kernel, we can perform PCA in the higher dimensional space. In this paper, we propose the kernel RPCA (KRPCA) and give its solution. Similarly to KPCA, the order of matrices that we should calculate for the solution is the number of samples, that is 'kernel trick'. We provide experimental results of an application to pattern recognition in order to show the advantages of KRPCA over KPCA.
研究論文（学術雑誌）, 英語 - Kernel sample space projection classifier for pattern recognition
Y Washizawa; Y Yamashita
PROCEEDINGS OF THE 17TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOL 2, IEEE COMPUTER SOC, 掲載ページ 435-438, 出版日 2004年, 査読付, We propose a new kernel-based method for pattern recognition. Support vector machine (SVM), principal component analysis (PCA), and Fisher discriminant have been extended to kernel based methods and they achieve better performance. In this paper we propose kernel sample space projection classifier (KSP) for pattern recognition. In KSP an unknown input pattern is discriminated by comparing the norms onto kernel sample spaces which are spanned by sample vectors mapped to a high dimensional feature space by Mercer kernel function.
In this paper we provide a closed form of our method and show its advantages by experimental results of the recognition problem using handwritten digit database "MNIST" and some two-class classification problems. Finally we compare it with other methods from several points of view.
研究論文（国際会議プロシーディングス）, 英語
MISC
- 符号変調視覚誘発電位のニューラルデコーディングと脳コンピュータインターフェースへの応用 (医用画像)
佐藤 純一; 鷲沢 嘉一
電子情報通信学会, 出版日 2016年05月19日, 電子情報通信学会技術研究報告 = IEICE technical report : 信学技報, 116巻, 39号, 掲載ページ 65-70, 日本語, 0913-5685, 40020849557, AA11370335 - GNU Octave (前編，後編)
鷲沢 嘉一
出版日 2011年, 映像情報メディア学会誌, 65巻, 5,6号, 掲載ページ 683-686, 790-793, 日本語, 記事・総説・解説・論説等（その他） - 多チャンネルの観測信号に含まれる共通信号の強度推定と抽出およびその脳電図処理への応用(音響信号処理,一般)
鷲沢 嘉一; 山下 幸彦; 田中 聡久; アンジェイ チホツキ
雑音が含まれている多チャンネルの信号から共通に含まれている信号の強度を推定し,抽出する手法を提案する.提案するモデルは信号源の数が観測数よりも1つ多いため,従来のブラインド信号分離では解くことが難しい.本論文では,共通信号の混合比を求めるために,交互最小二乗法などを用いたアルゴリズムを提案する.共通信号の混合比を求めることにより,1)共通信号抽出する.2)共通信号の強度やSNRの高いセンサーを見つける.3)共通信号の分布図を求める.などのことが可能となる.共通信号を抽出するために2つの評価基準を導入し,これを満たす解を求め,2つの解が同値であることを示す.また,人工データと脳電図(EEC)を用いた例を示し,提案アルゴリズムの有効性を示す., 社団法人電子情報通信学会, 出版日 2007年05月17日, 電子情報通信学会技術研究報告. EA, 応用音響, 107巻, 62号, 掲載ページ 61-66, 英語, 0913-5685, 110006291566
書籍等出版物
- 工学のためのフーリエ解析 (工学のための数学)
山下幸彦; 田中聡久; 鷲沢嘉一
学術書, 日本語, 共著, 数理工学社, 出版日 2016年 - 機械学習によるパターン識別と画像認識への応用 ～Octave/MATLABシミュレーションでわかりやすく解説～
鷲沢嘉一; 田中聡久
日本語, 共著, トリケップス, 出版日 2013年05月 - Principal Component Analysis
Parinya Sanguansat; Cuauhtemoc Araujo-Andrade; Claudio Frausto-Reyes; Esteban Gerbino; Pablo Mobili; Elizabeth Tymczyszyn; Edgar L. Esparza-Ibarra; Rumen Ivanov-Tsonchev; Andrea Go mez-Zavaglia; Maria Monfreda; Yoshikazu Washizawa; Xian-Hua Han; Yen-Wei Chen; Ramana Vinjamuri; WeiWang; Mingui Sun; Zhi-Hong Mao; Masahiro Kuroda; Yuichi Mori; Masaya Iizuka; Michio Sakakihara; Mauridhi Hery Purnomo; Diah P. Wulandari; I. Ketut; Eddy Purnama; Arif Muntasa; Shiow-Jyu Lin; Kun
英語, 共著, Chapter 4, Subset Basis Approximation of Kernel Principal Component Analysis, Intech, 出版日 2012年03月
講演・口頭発表等
- 相関分析を用いた CNN による符号変調 BCI
小谷 裕也; 鷲沢 嘉一
第39回信号処理シンポジウム
発表日 2024年12月
開催期間 2024年12月 - ヒルベルト変換を用いた複素CNNによるBCI
高田 倫太朗; 鷲沢 嘉一
日本語, 第39回信号処理シンポジウム
発表日 2024年12月
開催期間 2024年12月 - 位相偏移変調を用いた定常状態視覚誘発電位BCI
前島 蒼大; 鷲沢 嘉一
口頭発表（一般）, 日本語, 第39回信号処理シンポジウム
発表日 2024年12月
開催期間 2024年12月 - うつ病予防のためのマインドフルネス瞑想の脳波による定量的評価
徳永 翼; 鷲沢 嘉一
第12回看護理工学会学術集会, 査読付
発表日 2024年11月
開催期間 2024年11月 - 正定値対称行列空間上におけるリーマングラフを用いた計量学習による脳コンピュータインターフェース
佐々木 陸行; 鷲沢 嘉一
口頭発表（一般）, 日本語, 第38回信号処理シンポジウム
発表日 2023年11月
開催期間 2023年11月06日- 2023年11月08日 - レーザ変位計と振動スピーカを用いた能動ノイズ制御
海老江 航太; 鷲沢 嘉一
口頭発表（一般）, 日本語, 第38回信号処理シンポジウム
発表日 2023年11月
開催期間 2023年11月- 2023年11月 - 恐怖感情に対するタッチング効果を脳波を用いて評価する～VR刺激を用いた国際比較～
日吉 和子; 徳永翼; 鷲沢嘉一; Francois Vialatte
第11回看護理工学会学術集会, 査読付
発表日 2023年06月10日
開催期間 2023年06月10日- 2023年06月11日 - 恐怖に対するタッチング効果～脳波を用いた定量的評価～
徳永翼; 鷲沢嘉一; Francois Vialatte; 日吉和子
口頭発表（一般）, 日本語, 第10回看護理工学会学術集会
発表日 2022年10月 - Feature selection using L2,2-1 regularized deep neural network
鄭 子徳; 鷲沢嘉一
口頭発表（一般）, 日本語, 第21回情報科学技術フォーラム (FIT2022), 国内会議
発表日 2022年09月15日 - 複素CNNによる脳コンピューターインターフェースのためのSSVEP分類
池田 晟; 鷲沢嘉一
口頭発表（一般）, 日本語, 電子情報通信学会総合大会2021, 電子情報通信学会, 国内会議
発表日 2021年03月 - AUC Maximization in Deep Neural Network Learning for Imbalanced Classification Problems
Shuyu Zhang; Yoshikazu Washizawa
口頭発表（一般）, 英語, 電子情報通信学会, 電子情報通信学会, 国内会議
発表日 2021年03月 - 複素ニューラルネットワークを用いた自発脳波分類
池田晟; 鷲沢嘉一
口頭発表（一般）, 日本語, 電子情報通信学会信号処理研究会シンポジウム, 国内会議
発表日 2019年 - 畳み込みニューラルネットワークを用いた乾式電極EEGの信号復元
小上馬悠貴; 鷲沢嘉一
口頭発表（一般）, 日本語, 計測自動制御学会 ライフエンジニアリング部門シンポジウム2018, 国内会議
発表日 2018年 - ネガティブな視覚認知に対するタッチング効果 ~脳波を用いた感情調節の定量的評価~
村松 敬司; 鷲沢 嘉一; 日吉 和子
口頭発表（一般）, 日本語, 看護理工学会, 国内会議
発表日 2018年 - 多クラスヒンジ損失を用いたスパース表現分類器
紙屋亮太; 鷲沢嘉一
口頭発表（一般）, 日本語, 電子情報通信学会信号処理研究会シンポジウム, 国内会議
発表日 2018年 - 脳波情報に基づく音楽推薦システム
糸賀弘樹; 鷲沢嘉一
口頭発表（一般）, 日本語, 電子情報通信学会信号処理研究会シンポジウム, 国内会議
発表日 2017年11月 - ニューラルネットワークを用いた自発脳活動解析による音楽ジャンルの推定
糸賀弘樹; 鷲沢嘉一
口頭発表（一般）, 日本語, 信学技報, 国内会議
発表日 2017年 - 脳波を用いた音楽鑑賞時の感情認識における有効な特徴量の検討
口頭発表（一般）, 日本語, 国内会議
発表日 2016年 - 符号変調視覚誘発電位のニューラルデコーディングと脳コンピュータインターフェースへの応用
佐藤純一; 鷲沢嘉一
口頭発表（一般）, 日本語, 信学技報SIP2016-13, 国内会議
発表日 2016年 - 符号変調視覚誘発電位脳コンピュータインターフェースのための信頼度に基づく自動再送要求の適用に関する研究
佐藤純一; 鷲沢嘉一
ポスター発表, 日本語, 信学技報IEICE-EA2014-97, 国内会議
発表日 2015年03月 - P300-N100 speller and short code modulated VEP brain computer interfaces,
Y. Washizawa
口頭発表（招待・特別）, 英語, The Third APSIPA Workshop on the Frontier in Biomedical Signal Processing and Systems, 招待, 上海, 国際会議
発表日 2015年 - 音楽認知における$\gamma$活動の意義-脳波脳磁場同時記録を用いて
浦上裕子; 川村光毅; 鷲沢嘉一; アンジェイチホツキ
口頭発表（一般）, 日本語, 第43回日本臨床神経生理学会学術大会
発表日 2014年11月 - N100 use in brain computer interfaces,
Y. Washizawa
口頭発表（招待・特別）, 英語, The Second APSIPA Workshop on the Frontier in Biomedical Signal Processing and Systems (BioSiPS 2014), 招待, APSIPA, 国際会議
発表日 2014年03月 - 部分空間を用いた脳波識別のための空間フィルタの設計
善本秀法; 鷲沢嘉一
口頭発表（一般）, 日本語, 電子情報通信学会信号処理研究会, 国内会議
発表日 2014年 - Grassmannian Representation for Variational Pattern Classification and Its Application to Brain Signal Processing
Y. Washizawa
口頭発表（招待・特別）, 英語, The First APSIPA Workshop on the Frontier in Biomedical Signal Processing and Systems, Asia Pacific Signal and Information Processing Association, Bangkok, Thailand, 国際会議
発表日 2013年03月 - ECT法による生体情報の取得に関する研究
玉井瑞又; 鷲沢嘉一; 三上直樹
口頭発表（一般）, 日本語, 生活生命支援医療福祉工学系学会連合大会,生活生命支援医療福祉工学系学会連合大会
発表日 2013年 - 音楽聴取時の脳内神経基盤の解明
浦上裕子; 川村光毅; 鷲沢嘉一; 日吉和子
口頭発表（一般）, 日本語, 第42回日本臨床神経生理学会学術大会,第42回日本臨床神経生理学会学術大会
発表日 2012年10月 - 音声刺激による聴覚ブレイン・コンピュータ・インタフェースの可能性
田村潤; 鷲沢嘉一; 東広志; 田中聡久
口頭発表（一般）, 日本語, 信学技報,電子情報通信学会信号処理研究会
発表日 2012年03月 - 重み付け更新を用いたカーネル部分空間法による多クラス識別
鷲沢 嘉一
シンポジウム・ワークショップパネル（公募）, 日本語, 画像の認識・理解シンポジウム(MIRU 2012)
発表日 2012年 - ノイズ除去手法
鷲沢嘉一
その他, 日本語, 技術講習会, 海洋音響学会
発表日 2011年12月 - Grassmannian上のMahalanobis距離とその拡張に関する検討
鷲沢 嘉一; 堀田 政二
口頭発表（一般）, 日本語, 電子情報通信学会信号処理研究会,第26回信号処理シンポジウム
発表日 2011年11月 - オンライン部分カーネル主成分分析
鷲沢 嘉一
口頭発表（一般）, 日本語, 信学技報
発表日 2011年 - 定常的聴覚誘発電位を用いた脳コンピュータインターフェイス
東 広志; 鷲沢 嘉一; T. Rutkowski; 田中 聡久
口頭発表（一般）, 日本語, 信学技報
発表日 2011年 - 中心化部分カーネル主成分分析によるデノイジング
鷲沢 嘉一; 田中 正行
口頭発表（一般）, 日本語, 第9回情報科学技術フォーラム(FIT2010)
発表日 2010年 - パターン認識技術の基礎と実装 --プログラムで理解するパターン認識の知って得するテクニック--
鷲沢 嘉一; 堀田 政二
口頭発表（招待・特別）, 日本語, 第15回画像センシングシンポジウム(SSII09)
発表日 2009年 - 抑制付きカーネル部分空間法
鷲沢 嘉一
シンポジウム・ワークショップパネル（公募）, 日本語, 部分空間法研究会2008
発表日 2008年 - 制約付きの近似による特徴抽出
鷲沢 嘉一
シンポジウム・ワークショップパネル（公募）, 日本語, 第11 回画像の認識・理解シンポジウム(MIRU2008)
発表日 2008年 - マルチクラス識別問題に対する最適識別境界の学習
鷲沢 嘉一; 堀田 政二
口頭発表（一般）, 日本語, 電子情報通信学会,2008年総合大会
発表日 2008年 - 雑音を含む観測信号からのブラインド信号抽出と信号処理への応用
鷲沢 嘉一; 山下 幸彦; 田中 聡久; A. Cichocki
口頭発表（一般）, 日本語, 電子情報通信学会,2008年総合大会
発表日 2008年 - マージン最大化基準に基づく特徴選択と脳信号処理への応用
鷲沢 嘉一
シンポジウム・ワークショップパネル（公募）, 日本語, 第23回信号処理シンポジウム講演論文集，, 電子情報通信学会
発表日 2008年 - 多チャンネルの観測信号に含まれる共通信号の強度推定と抽出およびその脳電図処理への応用
鷲沢 嘉一; 山下 幸彦; 田中 聡久; A. Cichocki
口頭発表（一般）, 日本語, 信学技報
発表日 2007年 - 正則化を用いた２次識別器
鷲沢 嘉一
シンポジウム・ワークショップパネル（公募）, 日本語, 画像の認識・理解シンポジウム(MIRU 2007)
発表日 2007年 - 経験的モード分解における帯域幅制御
鷲沢 嘉一; 田中 聡久; D. Mandic; A. Cichocki
シンポジウム・ワークショップパネル（公募）, 日本語, 第21 回信号処理シンポジウム講演論文集
発表日 2007年 - オンラインK 平面クラスタリングによる疎行列解析
鷲沢 嘉一; A. Cichocki
口頭発表（一般）, 日本語, 2006 年電子情報通信学会総合大会講演論文集,2006 年電子情報通信学会総合大会講演論文集
発表日 2006年 - A family of kernel subspace classifiers
鷲沢 嘉一; 山下 幸彦
口頭発表（一般）, 英語, 部分空間法研究会2006 予稿集，,部分空間法研究会2006 予稿集，
発表日 2006年 - 抑制型カーネル標本空間射影法によるパターン認識
鷲沢 嘉一; 山下 幸彦
口頭発表（一般）, 日本語, 信学技報,信学技報
発表日 2004年 - カーネルウィーナーフィルタ
鷲沢 嘉一; 山下 幸彦
口頭発表（一般）, 日本語, 信学技報,信学技報
発表日 2004年 - カーネル相対主成分分析による 多クラスパターン認識
鷲沢 嘉一; 疋田謙司; 田中聡久; 山下幸彦
口頭発表（一般）, 日本語, 第2 回 FIT（情報科学技術フォーラム）情報レターズ,第2 回 FIT（情報科学技術フォーラム）情報レターズ
発表日 2003年 - カーネルウィーナーフィルタ
鷲沢 嘉一; 山下 幸彦
シンポジウム・ワークショップパネル（公募）, 日本語, 第6 回情報論的学習理 論ワークショップ
発表日 2003年 - カーネル標本空間射影法によるパターン認識
鷲沢 嘉一; 山下 幸彦
シンポジウム・ワークショップパネル（公募）, 日本語, 第6回情報論的学習理論ワークショップ
発表日 2003年 - パターン認識のための相対KL 変換法の高精度化,
鷲沢 嘉一; 疋田謙司; 田中 聡久; 山下幸彦
口頭発表（一般）, 日本語, 第１回 FIT （情報科学技術フォーラム） 一般講演論文集
発表日 2002年
担当経験のある科目_授業
- 数理統計（Mエリア）
現在
電気通信大学 - 電子工学工房
現在
電気通信大学 - 数理統計（Mエリア）
The University of Electro-Communications - 電子情報学実験A
The University of Electro-Communications - 電子情報学実験A
電気通信大学 - 輪講Ａ（学部・Ｋ課程）
The University of Electro-Communications - 輪講Ａ（学部・Ｋ課程）
電気通信大学 - ディジタル信号処理基礎論
電気通信大学 - イノベイティブ総合コミュニケーションデザイン1,2
電気通信大学 - イノベイティブ総合コミュニケーションデザイン1,2
電気通信大学 - 電気回路学および演習
電気通信大学 - 電気回路学および演習
電気通信大学 - ディジタル信号処理基礎論
The University of Electro-Communications - ディジタル信号処理基礎論
電気通信大学 - 情報・通信工学基礎
電気通信大学 - 情報・通信工学基礎
電気通信大学 - 電子工学工房
The University of Electro-Communications
共同研究・競争的資金等の研究課題
- 認識機構のファイバー束による統一的表現理論の構築とその機械学習への応用
研究期間 2020年04月01日 - 2023年03月31日 - 介護の見える化：タッチングケア効果の評価
公益財団法人KDDI財団
研究期間 2019年 - 2022年 - 作用素多様体理論の構築とパターン認識への応用
山下幸彦
研究期間 2017年04月01日 - 2020年03月31日 - カーネルグラスマン表現の計量構造及び脳信号処理への応用
鷲沢嘉一
研究代表者
研究期間 2015年04月01日 - 2018年03月31日 - 脳波・脳磁図と機能的MRIを用いた脳損傷者の安静時機能的脳活動の解明
浦上裕子
研究期間 2015年04月01日 - 2018年03月31日
