Language：English   Publishing type：Research paper (scientific journal)

Forgetting is important for animals to manage acquired memories to enable adaptation to changing environments; however, the neural network in mechanisms of forgetting is not fully understood. To understand the mechanisms underlying forgetting, we examined olfactory adaptation, a form of associative learning, in Caenorhabditis elegans (C. elegans). The forgetting of diacetyl olfactory adaptation in C. elegans is regulated by secreted signals from AWC sensory neurons via the TIR-1/JNK-1 pathway. These signals cause a decline of the sensory memory trace in AWA neurons where diacetyl is mainly sensed. To further understand the neural network that regulates this forgetting, we investigated the function of interneurons downstream of AWA and AWC neurons. We found that a pair of interneurons, AIA, is indispensable for the proper regulation of behavioral forgetting of diacetyl olfactory adaptation. Loss of or inactivation of AIA caused the impairment of the chemotaxis recovery after adaptation without causing severe chemotaxis defects in naïve animal. AWA Ca2+ imaging analyses suggested that loss or inactivation of AIA interneurons did not affect the decline of the sensory memory trace after the recovery. Furthermore, AIA responses to diacetyl were observed in naïve and after the recovery, but not just after the conditioning, suggesting that AIA responses after the recovery are required for the chemotaxis to diacetyl. We propose that the functional neuronal circuit for attractive chemotaxis to diacetyl is changed temporally at the recovery phase so that AIA interneurons are required for chemotaxis, although AIAs are dispensable for attractive chemotaxis to diacetyl in naïve animals.

DOI： https://doi.org/10.1523/ENEURO.0084-22.2022

DOI： https://doi.org/10.1523/ENEURO.0084-22.2022

Other Link： https://www.eneuro.org/content/early/2022/08/17/ENEURO.0084-22.2022

Authorship：Last author,　Corresponding author   Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)

日本   姫路   2022.07.25  -  2022.07.28

本稿は半教師あり二値分類のための新しいコスト関数を 提案する. 提案するコスト関数は, Focal Loss を半教師あり分類のために拡張した損失関数とクラス事前確率罰則項 から構成される. クラス事前確率罰則項を導入することで, ラベルなしデータにおける正例の割合がわかる場合, その 事前情報を直接的に推論に活用することが可能である. 提 案するコスト関数の有用性を評価するため, CIFAR-10 を 用いた画像二値分類試験を行った. その結果, 少量のラベ ルありデータセットや不均衡なデータセットにおいて, 提 案するコスト関数は既存のコスト関数と比べて優れた性能を示すことを確認した.

Other Link： https://sites.google.com/view/miru2022/program#h.44rbhnwbmvlj

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1523/ENEURO.0084-22.2022

Kyutacar

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85137058885&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

We propose an application of the independent component analysis (ICA) to separate satellite-induced time-varying stray fields from magnetic field data obtained using onboard multiple magnetometers. The ICA is a method for estimating source signals at multiple sites so that the estimated source signals can become statistically independent of each other. Since stray field variations are statistically independent of external natural field variations, the ICA method is expected to separate the natural variations from stray fields. Thus, we applied the ICA to magnetic field data from the first Quasi-Zenith Satellite, which has two triaxial fluxgate magnetometers, without using an extendable boom. First, we removed the long-period trend from the original data to create detrended data. Then, we applied the FastICA algorithm to the detrended data and obtained six independent components (ICs). The stray fields were successfully separated into three ICs (noise ICs), and the natural signals were represented by the other three ICs (signal ICs). Finally, we restored the observed signals from the signal ICs, and confirmed that the natural phenomena variations were not altered by the processing step. We also proposed a selection method of the noise ICs using the C coefficient, which is the coefficient of the variance of the mixing vectors. There was a large difference in C between the ICs whose C coefficients are the largest 3rd and 4th ones. Overall, these results demonstrate the possibility that the ICA method can support for boom-less magnetic observations in future satellite missions.

DOI： 10.1029/2020JA028790

Scopus

Other Link： https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020JA028790

Authorship：Last author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)

Virtual   Virtual   2021.04.28  -  2021.04.30

This paper proposes an unsupervised anomaly detection technique for image-based plant disease diagnosis. The construction of large and publicly available datasets containing labeled images of healthy and diseased crop plants led to growing interest in computer vision techniques for automatic plant disease diagnosis. Although supervised image classifiers based on deep learning can be a powerful tool for plant disease diagnosis, they require a huge amount of labeled data. The data mining technique of anomaly detection includes unsupervised approaches that do not require rare samples for training classifiers. We propose an unsupervised anomaly detection technique for image-based plant disease diagnosis that is based on the reconstructability of colors; a deep encoder-decoder network trained to reconstruct the colors of healthy plant images should fail to reconstruct colors of symptomatic regions. Our proposed method includes a new image-based framework for plant disease detection that u tilizes a conditional adversarial network called pix2pix and a new anomaly score based on CIEDE2000 color difference. Experiments with PlantVillage dataset demonstrated the superiority of our proposed method compared to an existing anomaly detector at identifying diseased crop images in terms of accuracy, interpretability and computational efficiency.

DOI： 10.5220/0010463201120120

DOI： 10.5220/0010463201120120

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115692116&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

Vascular endothelial cells (ECs) in angiogenesis exhibit inhomogeneous collective migration called “cell mixing”, in which cells change their relative positions by overtaking each other. However, how such complex EC dynamics lead to the formation of highly ordered branching structures remains largely unknown. To uncover hidden laws of integration driving angiogenic morphogenesis, we analyzed EC behaviors in an in vitro angiogenic sprouting assay using mouse aortic explants in combination with mathematical modeling. Time-lapse imaging of sprouts extended from EC sheets around tissue explants showed directional cohesive EC movements with frequent U-turns, which often coupled with tip cell overtaking. Imaging of isolated branches deprived of basal cell sheets revealed a requirement of a constant supply of immigrating cells for ECs to branch forward. Anisotropic attractive forces between neighboring cells passing each other were likely to underlie these EC motility patterns, as evidenced by an experimentally validated mathematical model. These results suggest that cohesive movements with anisotropic cell-to-cell interactions characterize the EC motility, which may drive branch elongation depending on a constant cell supply. The present findings provide novel insights into a cell motility-based understanding of angiogenic morphogenesis.

DOI： 10.1038/s41598-019-45666-2

Scopus

CiNii Article

CiNii Research

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85068037379&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1109/TCBB.2017.2782255

DOI： 10.1109/TCBB.2017.2782255

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85038382107&origin=inward

Language：Japanese   Publishing type：Research paper (bulletin of university, research institution)

DOI： 10.11316/jpsgaiyo.72.1.0_2817

CiNii Article

CiNii Research

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1371/journal.pcbi.1004970

DOI： 10.1371/journal.pcbi.1004970

Scopus

PubMed

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84978903985&origin=inward

Authorship：Lead author   Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1093/bioinformatics/btu271

DOI： 10.1093/bioinformatics/btu271

Scopus

PubMed

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84902509846&origin=inward

Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)

<p>細胞動画像の典型的な特徴として多数の追跡対象物が密に存在し視覚的に類似していることが挙げられる．このような動画像に対し標準的な粒子フィルタを利用した場合，本来の追跡対象を別の対象物と誤認し追跡に容易に失敗する．この問題に対し多数の細胞の共変動性に着目し追跡のための補助情報としてこれを利用することで追跡精度の向上を目指す．</p>

DOI： 10.11517/pjsai.jsai2014.0_2c32in

CiNii Article

CiNii Research

Authorship：Last author   Language：English   Publishing type：Research paper (international conference proceedings)

In this paper, we propose the algorithm to extract the groove feeling from drum data. In the previous researches, extraction of the groove feeling requires pre-separated acoustic sources. We employed non-negative matrix factorization (NMF) to make separated information on hitting time from monaural wave data in which multiple acoustic sources are mixed. We applied our algorithm to a drum data and obtained a difference of time fluctuations among instruments, which relates to groove feeling and impression. The result implies that the proposed algorithm can extract some sensitive differences of nuance of drumming. © 2012 IEEE.

DOI： 10.1109/SCIS-ISIS.2012.6505312

DOI： 10.1109/SCIS-ISIS.2012.6505312

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84877794120&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

Non-stationary effects are ubiquitous in real world data. In many settings, the observed signals are a mixture of underlying station ary and non-stationary sources that cannot be measured directly. For example, in EEG analysis, electrodes on the scalp record the activity from several sources located inside the brain, which one could only measure invasively. Discerning stationary and non-stationary contributions is an important step towards uncovering the mechanisms of the data generating system. To that end, in Stationary Subspace Analysis (SSA), the observed signal is modeled as a linear superposition of stationary and non-stationary sources, where the aim is to separate the two groups in the mixture. In this paper, we propose the first SSA algorithm that has a closed form solution. The novel method, Analytic SSA (ASSA), is more than 100 times faster than the state-of-the-art, numerically stable, and guaranteed to be optimal when the covariance between stationary and non-stationary sources is time-constant. In numerical simulations on wide range of settings, we show that our method yields superior results, even for signals with time-varying group-wise covariance. In an application to geophysical data analysis, ASSA extracts meaningful components that shed new light on the Pi 2 pulsations of the geomagnetic field. © 2012 Elsevier Ltd.

DOI： 10.1016/j.neunet.2012.04.001

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84863869958&origin=inward

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)

To predict an occurrence of extraordinary phenomena, such as earthquakes, failures of engineering systems and financial market crushes, it is important to identify precursory events in time series. However, existing methods are limited in their applicability for real world precursor detections. Recently, Ide and Inoue [1] have developed an SSA-based change-point detection method, called singular spectrum transformation (SST). SST is suitable for detecting various types of change-points, but real world precursor detections can be far more difficult than expected. In general, precursory events are observed as minute and less-visible fluctuations preceding an onset of massive fluctuations of extraordinary phenomena and therefore they are easily over-looked. To overcome this point, we extend the conventional SST to the multivariable SST. The originality of our strategy is in focusing on synchronism detections of precursory events in multiple sequences of univariate time series. We performed some experiments by using artificial data and showed the superiority of multivariable SST in detecting onset of precursory events. Furthermore, the superiority is also shown statistically in determining the onset of precursory events by using real world time series.

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80053970902&origin=inward

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (bulletin of university, research institution)

Pi 2 magnetic pulsations observed on the ground are a good indicator of the auroral breakup. However, they have not only corresponding full-substorm onsets but also most pseudobreakups. Another well-known substorm related phenomenon observed on the ground is positive bays. In order to identify full-substorm onsets from ground-magnetometer data, we developed a new algorithm based on "Singular Spectrum Analysis(SSA)". The algorithm enables us to screen Pi 2 pulsations accompanied by the magnetic positive bay. We applied proposed algorithm to ground-magnetometer data and compared to the obtained results with Polar/UVI data. As a result, we succeeded in identifying 62% of the full-substorm onsets from ground-magnetometer data obtained in the nighttime sector between 21 and 03LT.

DOI： 10.5109/19197

CiNii Article

CiNii Research

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1029/2010JA016042

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79959755995&origin=inward

Authorship：Lead author,　Corresponding author   Language：Japanese   Publishing type：Research paper (international conference proceedings)

一般に，前兆現象は突発現象にそのものに比べて非常に目立ちにくく，その開始時刻は曖昧である．従来よく用いられてきた変化点検出法を適用した場合，このような微小で緩慢な変化は見逃されやすい．Tokunaga et al.1) では，Ide and Inoue2) の提案した特異スペクトル分析を応用した変化点検出法 (SST) を，多次元データを用いたアルゴリズム (MSST) へと拡張することで，鋭敏に前兆現象の開始時刻を推定出来ることを示した．MSST は，緩慢な変化も検出できる鋭敏な手法であるが，実データへの適用では誤検出が問題になる．本稿では，突発現象の大まかな開始時刻を予め検出し，さらに検出された時刻の前後で前兆現象の開始時刻と終了時刻を個別に探索することで，誤検出を劇的に減少させることができることを示す．

CiNii Article

CiNii Research

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79958702189&origin=inward

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)

CiNii Research

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)

CiNii Research

Language：English   Publishing type：Research paper (scientific journal)

The objective of this study is to understand better the propagation of Pi 2 waves in the nighttime region. We examined Pi 2 oscillations that showed high correlation between high- and low-latitude Magnetic Data Acquisition System/Circum Pan-Pacific Magnetometer Network stations (correlation coefficient: |γ| ≥ 0.75). For each horizontal component (H and D) we examined the magnetic local time (MLT) dependence of the delay time of high-latitude Pi 2 oscillations that corresponds to the highest correlation with the low-latitude Pi 2 oscillation. We found the delay time of the high-latitude H showed remarkable MLT dependence, especially in the premidnight sector: we found that in the premidnight sector the high-latitude H oscillation tends to delay from the low-latitude oscillation (<100 s). On the other hand, the delay time of the high-latitude D oscillation was not significant (∼±10 s) in the entire nighttime sector. We propose a Pi 2 propagation model to explain the observed delay time of high-correlation highlatitude H. The model quantitatively explains the trend of the event distribution. We also examined the spatial distribution of high-correlation Pi 2 events relative to the center of auroral breakups. It was found that the high-correlation Pi 2 events tend to occur away from the center of auroral breakups by more than 1.5 MLT. The present result suggests that the high-correlation H component Pi 2 oscillations at high latitude are a manifestation of forced Alfvén waves excited by fast magnetosonic waves. Copyright 2009 by the American Geophysical Union.

DOI： 10.1029/2009JA014163

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=72149099064&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1186/BF03352828

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=52149085015&origin=inward

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.1029/2007GL030174

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=34548580614&origin=inward

Language：English   Publishing type：Article, review, commentary, editorial, etc. (other)

DOI： 10.1002/essoar.10512276.1

DOI： 10.1002/essoar.10512276.1

Language：English   Publishing type：Research paper (scientific journal)

Japan   オンライン  

This study evaluated the impact of Atmospheric River (AR) clouds on snowfall amounts based on limited observation data to estimate the surface mass balance (SMB) of Antarctica. To accomplish this, we attempted to identify the snowfall cloud at Syowa Station, Antarctica. We constructed a new convolutional neural network (CNN) architecture with multinomial and binary classifications for National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) images over five years. The CNN was based on VGG16, and concatenate layers were added as the inception module. We replaced all the convolution layers with global average pooling to reduce the number of parameters. Based on the positive CNN sample re- sult, the multinomial classification emphasized the entire cloud structure, while the binary classification focused on cloud continuity. The results indicated accuracies of 71.00% and 65.37% for binary and multinomial classifications, respectively.

DOI： 10.1007/978-3-030-73113-7_7

Other Link： https://link.springer.com/chapter/10.1007/978-3-030-73113-7_7

Authorship：Last author   Language：English   Publishing type：Research paper (scientific journal)

<p>This study evaluated snowfall values based on limited observation data to estimate the surface mass balance (SMB) of Antarctica. To accomplish this, we attempted to identify the snowfall cloud at Syowa Station, Antarctica. We constructed a new convolutional neural network (CNN) architecture with multinomial and binary classifications and added National Oceanic and Atmospheric Administration (NOAA)/Advanced Very High Resolution Radiometer (AVHRR) images over five years. The CNN was based on VGG16, and concatenate layers were added as the inception module. We replaced all the convolution layers with global average pooling to reduce the number of parameters. Based on the positive CNN sample result, the multinomial classification emphasized the entire cloud structure, while the binary classification focused on cloud continuity. The results indicated accuracies of 71.00% for binary and 65.37% for multinomial classifications.</p>

DOI： 10.11517/pjsai.JSAI2020.0_3F1ES205

CiNii Article

CiNii Research

Other Link： https://ci.nii.ac.jp/naid/130007857079

Authorship：Lead author,　Last author,　Corresponding author   Language：Japanese   Publishing type：Article, review, commentary, editorial, etc. (scientific journal)

DOI： 10.11499/sicejl.58.166

CiNii Article

CiNii Research

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I029600573

Language：English   Publishing type：Research paper (conference, symposium, etc.)

DOI： 10.14851/jcsir.2019.0_119

CiNii Article

CiNii Research

Other Link： https://ci.nii.ac.jp/naid/130007743208

Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)

DOI： 10.11316/jpsgaiyo.73.2.0_2315

CiNii Article

CiNii Research

Language：Japanese   Publishing type：Research paper (international conference proceedings)

DOI： 10.11316/jpsgaiyo.73.1.0_2853

CiNii Article

CiNii Research

Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)

日本   大阪   2017.03.17  -  2017.03.20

DOI： https://doi.org/10.11316/jpsgaiyo.72.1.0_2817

Language：Japanese   Publishing type：Research paper (bulletin of university, research institution)

CiNii Article

CiNii Research

Other Link： http://id.ndl.go.jp/bib/026589082

Authorship：Corresponding author   Language：Japanese   Publishing type：Article, review, commentary, editorial, etc. (other)

J-GLOBAL

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.5047/eps.2011.10.013

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84876370337&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

DOI： 10.5194/angeo-29-1663-2011

Scopus

Other Link： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80053422182&origin=inward

Language：English   Publishing type：Research paper (scientific journal)

DOI： https://doi.org/10.5194/angeo-29-1663-2011

Authorship：Lead author,　Corresponding author   Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)

第6回宇宙環境シンポジウム (2009年2月29日-30日. 北九州国際会議場)

CiNii Article

CiNii Research

Other Link： http://id.nii.ac.jp/1696/00005141/

Authorship：Lead author,　Corresponding author   Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)

CiNii Research

Event date： 2024.08.06 - 2024.08.09   Language：Japanese   Country：Japan  

Other Link： https://miru-committee.github.io/miru2024/

Event date： 2024.08.06 - 2024.08.09   Language：Japanese   Country：Japan  

Other Link： https://miru-committee.github.io/miru2024/

Event date： 2024.08.06 - 2024.08.09   Language：Japanese   Country：Japan  

Other Link： https://miru-committee.github.io/miru2024/

Event date： 2024.08.06 - 2024.08.09   Language：Japanese   Country：Japan  

Other Link： https://miru-committee.github.io/miru2024/

Event date： 2024.05.28 - 2024.05.31   Language：Japanese   Country：Japan  

Event date： 2024.03.01 - 2024.03.02   Language：English   Country：Japan  

Other Link： https://www.brain.kyutech.ac.jp/~neuro/2023/11/30/5th-sympo/?lang=en

Event date： 2023.11.14 - 2023.11.15   Language：English   Country：Japan  

Other Link： https://www.nipr.ac.jp/symposium2023/

Event date： 2023.09.07 - 2023.09.08   Language：Japanese   Country：Japan  

Other Link： https://smartconf.jp/content/iibmp2023/posterpresentation

Event date： 2023.09.07 - 2023.09.08   Language：Japanese   Country：Japan  

Other Link： https://smartconf.jp/content/iibmp2023/posterpresentation

Event date： 2023.07.26 - 2023.07.29   Language：Japanese   Country：Japan  

Event date： 2023.07.26 - 2023.07.29   Language：Japanese   Country：Japan  

Event date： 2023.05.21 - 2023.05.26   Language：English   Country：Japan  

Event date： 2022.12.19 - 2022.12.21   Language：Japanese   Country：Japan  

Event date： 2022.12.13 - 2022.12.14   Language：English   Country：Japan  

Semiconductor visual inspection systems are required to further improve productivity and quality control in proportion to the increasing demand for semiconductor products. In machine vision systems, the combination of conventional rule-based anomaly detection methods, which are characterized by their high explanatory power for the basis of judgment, and deep learning methods, which are characterized by their high extraction and classification accuracy for features such as complex shapes, is an effective means from the perspective of productivity and quality assurance.
We propose a rule-based anomaly detection method that uses binary segmentation [1] as a preprocessing step. Figure 1 shows the inspection flow model of the proposed method. Binary segmentation is applied to the input image for regions that are difficult to distinguish from defect regions by clear human definition and that are a cause of over-detection with characteristics such as shape and pixel value. The extracted regions are then combined with the input image as a mask. Then, a rule-based abnormality detection inspection is applied to the synthesized image.

Event date： 2022.12.13 - 2022.12.14   Language：English   Country：Japan  

Event date： 2022.12.13 - 2022.12.14   Language：English   Country：Japan  

The nervous system of Caenorhabditis elegans (C. elegans) realizes sensory-motor transformation for many odorants despite it consists of only 302 neurons. Recently, we have developed a novel live-cell imaging technique based on genetically encoded voltage and calcium indicators. Our imaging system enables simultaneous measurements of voltage and Ca2+ responses to specific sequential odor stimulation at a single cellular level. The series of experiments interestingly suggest that an individual neuron can sensitively and quantitatively express time-varying environmental information quite unlike the formal neuron. In this study, we attempted to model a single-neuron of C. elegans in a data-driven manner. As a pilot study, we validated whether voltage responses in AWA, which is one of the olfactory sensory neurons of C. elegans, can be reproduced by a simple echo state network (ESN). For input data, we used synthetic square waves that imitate sequential odor stimulation in in-vivo experiments. For target signals, actual measurements of voltage responses in AWA were used. The prediction test by the trained ESN showed that the transient variations of membrane potentials can be successfully reproduced while we failed to reproduce quasi-stationary depolarizations during odorant stimulation (Figure 1). The result warrants further research on more realistic ESN models including multiple voltage dependent ion channels.

Event date： 2022.12.13 - 2022.12.14   Language：English   Country：Japan  

Supervised learning requires a large amount of labeled data to obtain high generalization performance. The labeled data is created by annotation. However, the annotation has a lot of problems, such as high cost. In this study, we propose a novel semi-supervised cost function for using unlabeled data effectively, which consists of two important terms. There is a class-prior probability penalty and an extended Focal loss function. A class-prior probability prevents overfitting to labeled data by utilizing prior knowledge for a percentage of positive samples over unlabeled data. We performed binary image segmentation to evaluate the proposed cost function's effectiveness. As a result, the proposed cost function performs flexible segmentation of objects with ambiguous shapes, because the output values of the boundary regions transition smoothly compared with Binary Crossentropy (BCE) and PNU Loss.

Event date： 2022.11.20 - 2022.11.23   Language：Japanese   Country：Japan  

Event date： 2022.11.20 - 2022.11.23   Language：Japanese   Country：Japan  

Event date： 2022.09.13 - 2022.09.15   Language：English   Country：Japan  

Event date： 2022.08.29 - 2022.08.30   Language：Japanese  

Event date： 2022.03.18 - 2022.03.19   Language：English   Country：Japan  

Event date： 2021.12.02 - 2021.12.08   Language：Japanese  

近年，水蒸気輸送と豪雨(雪)には対流性の雲が連なって形成されるAtmospheric River(AR)が関連していると考えられるようになり，南極域においても極方向の水蒸気フラックス強化となる背景場とARの観測事例が報告されている[1]．そこで，昭和基地の降雪時にARと判別できる雲画像を用いたCNNによる自動識別に取り組んできた[2]．また，客観解析データを用いた全球規模のARとエアロゾル輸送の関連について報告がされており，Aerosol Atmospheric River (AAR)となって高濃度エアロゾルが輸送されることが示されている．本研究では，これまで行ってきたARによる水蒸気輸送だけではなく，陸起源と考えられる大気中微量物質の輸送も同時に捉え，大気による物質輸送過程を機械学習による予測モデルを構築することを目指す．まずは，ARとエアロゾル輸送の関連について2009年のブリザードイベントについて調べた．

Event date： 2021.11.10 - 2021.11.13   Language：Japanese  

Event date： 2021.11.10 - 2021.11.13   Language：Japanese  

Event date： 2021.11.10 - 2021.11.13   Language：Japanese  

Event date： 2021.11.10 - 2021.11.13   Language：Japanese  

Event date： 2021.10.31 - 2021.11.04   Language：Japanese  

Event date： 2021.09.21 - 2021.09.23   Language：English  

C. elegans is a model organism in which the structure (connectome) of the whole nervous system composed of 302 neurons has been determined. We performed whole-brain imaging by spinning disk confocal microscope combined with piezo objective positioner to obtain calcium imaging data of the whole head neurons. Further, we modeled the dynamics of neuronal ensembles based on the observed activity data and connectome data. As a result, we could perform virtual ablation of neurons or particular connections to gain insights into the information flow through the neural circuits.

Event date： 2021.06.21 - 2021.06.24   Language：English  

Measurement of neuronal activities in non-invasive and unanesthetized condition is important for understanding neuronal function in intact animals. Ca2+ imaging by fluorescent gene encoded calcium indicators (GECI) are a powerful way to measure neuronal activities in C. elegans. Although Ca2+ imaging revealed important aspects in neuronal functions, the measurement of neuronal membrane voltage is important to understand the neuronal functions. Furthermore, the relations of change of membrane voltages and changes of Ca2+ has not been fully understood. Recently, several types of gene encoded voltage indicators (GEVI) that are derived from 7TM proteins used for optogenetics has been developed to measure changes of membrane voltage in living animals. Even though the fluorescence of these GEVIs is dim, they showed fast time constants and relatively high fluorescent change depend on voltages. Among those GEVIs, we use paQuasAr3 for the voltage measurement, because it shows relatively higher fluorescence with other superior characteristics.
Since AWA, one of the olfactory sensory neurons, which is responsible for diacetyl sensation, was reported to show all-or-none action potentials (Liu et al. 2018), we firstly analyzed AWA voltage changes induced by diacetyl. We found that fluorescence of paQuasAr3 expressed in AWA cell body is changed in response to diacetyl stimulation with high reproducibility. At the beginning of the stimulation, the transient increase and decrease of fluorescence intensity was observed, whereas the relatively higher fluorescence intensity was sustained during the stimulation. To elucidate relations between the Ca2+ responses and the voltage responses, we made wild-type animals expressing paQuasAr3 and GCaMP6f in AWA neurons, and measured both fluorescence at a cell body simultaneously. We found that the changes of paQuasAr3 started faster than the changes of GCaMP. These analyses will give insights on the neuronal functions in informational processing.

Event date： 2021.05.30 - 2021.06.06   Language：Japanese  

Event date： 2021.05.10   Language：Japanese  

Event date： 2021.04.28 - 2021.04.30   Language：English  

This paper proposes an unsupervised anomaly detection technique for image-based plant disease diagnosis. The construction of large and publicly available datasets containing labeled images of healthy and diseased crop plants led to growing interest in computer vision techniques for automatic plant disease diagnosis. Al- though supervised image classifiers based on deep learning can be a powerful tool for plant disease diagnosis, they require a huge amount of labeled data. The data mining technique of anomaly detection includes un- supervised approaches that do not require rare samples for training classifiers. We propose an unsupervised anomaly detection technique for image-based plant disease diagnosis that is based on the reconstructability of colors; a deep encoder-decoder network trained to reconstruct the colors of healthy plant images should fail to reconstruct colors of symptomatic regions. Our proposed method includes a new image-based framework for plant disease detection that utilizes a conditional adversarial network called pix2pix and a new anomaly score based on CIEDE2000 color difference. Experiments with PlantVillage dataset demonstrated the superiority of our proposed method compared to an existing anomaly detector at identifying diseased crop images in terms of accuracy, interpretability and computational efficiency.

Event date： 2021.03.25   Language：Japanese  

Event date： 2021.03.12 - 2021.03.15   Language：Japanese  

Event date： 2020.12.08   Language：Japanese  

Event date： 2020.12.02 - 2020.12.04  

Event date： 2020.11.23 - 2020.11.26   Language：English  

Event date： 2020.11.23 - 2020.11.26   Language：Japanese  

Event date： 2020.11.01 - 2020.11.04   Language：English  

The machine learning has become a powerful tool to find the relation between variables thanks to the deep learning technique. This performs greatly in the classification, regression and recently generative modeling in the engineering and commercial areas. However, due to the satisfaction of physical laws in the scientific research area, the application of machine learning has some difficulties. In particular, the generative modeling is very sensitive to scientific data since the generated data is not guaranteed by the physical laws.
To overcome these problems, we have tried to apply machine learning to space plasma physics. In the observation there are many lacks data in space and time. Using the technique of GAN (Generative Adversarial Networks), we have challenged to represent the lack data of aurora image by ASI (All-Sky Imager) of THEMIS. Now we use the natural training data not only the observation data and we have obtained the smooth represented data, however these data cannot satisfy the physical laws. Then we prepare the training data of only observation.
From this thought the preparing the training data is the most important for machine learning. Then we have prepared the global simulation data of magnetosphere using real solar wind data for the generation and forecast the configuration of the magnetosphere. These data are the very large size and time elapsed data so that the data set cannot be stored in often case and usual machine learning cannot treat these data set. However recently there are 3D CNN (convolutional neural network) and RNN (recurrent neural network) which can be trained by 3D data set and these data set may become very important. In this study, we show the database of this data set data, representation of the auroral image and their status.

Event date： 2020.09.01 - 2020.09.03   Language：Japanese  

Event date： 2020.08.02 - 2020.08.05   Language：Japanese  

本論文では, 色に現れる異常の検出を目的とした, 教師なし異常検出方法を提案する. 提案手法では, 敵対的生成ネットワークに基づく教師なし学習に基づき, 色の再構成可能性を評価することで, 正常データでは見られなかった色の異常の検知を行う. また, 色の再構成可能性の評価のため, CIEDE2000 色差に基づく異常度スコアを提案する. 実験では, PlantVillage データセットを用いた病気の植物とその病変領域の検出に対して提案手法の性能評価を行い,ベースラインである AnoGAN との比較を行う. 実験の結果, 提案手法が色の異常検知問題に対して検出性能, 解釈可能性および計算効率性の観点から AnoGAN より優れた性能を発揮することを示す.

Event date： 2020.08.02 - 2020.08.05   Language：Japanese  

Event date： 2020.07.12 - 2020.07.16   Language：English  

We propose an application of the independent component analysis (ICA) to separate satellite- induced time-varying noises from magnetic data obtained by onboard multiple magnetometers. The ICA is a method of estimating source signals observed at multiple sites so that estimated source signals are statistically independent of each other. Since satellite noises are clearly independent of external natural variations, the ICA is expected to separate the satellite noises. We applied the ICA to the magnetic data measured by the first Quasi-Zenith Satellite (QZS-1) that has two triaxial fluxgate magnetometers without an extendable mast. First, we removed the long-period trend from the original data. Then, we applied the FastICA to the detrended data and obtained six independent components (ICs). Satellite-induced noises were successfully classified into three ICs (noise ICs). Natural signals are represented by the rest three ICs (signal ICs). Finally, we restored the observed signals from the signal ICs. We confirmed that amplitude and waveform of natural phenomena were not altered by the processing. We also proposed the method of automatic determination of noise ICs using the D score, which is similar to the normalized coefficient of variance of the mixing vectors. We confirmed that the three largest D scores, which give the noise ICs, are much larger than the three smallest D, which give the signal ICs. The automatic determination of noise ICs by this method was 95% identical to that by visual inspection. These results demonstrated that the ICA method can provide for mast-less magnetic observations in future satellite missions.

Event date： 2020.06.09 - 2020.06.12   Language：English  

Event date： 2020.05.24 - 2020.05.28   Language：English  

Event date： 2020.02.28   Language：Japanese  

Event date： 2019.12.18 - 2019.12.20   Language：Japanese  

Event date： 2019.12.18 - 2019.12.20   Language：Japanese  

Event date： 2019.12.09 - 2019.12.13   Language：English  

Event date： 2019.12.03 - 2019.12.06   Language：Japanese  

Event date： 2019.12.03 - 2019.12.06   Language：Japanese  

Event date： 2019.11.01 - 2019.11.05   Language：English  

Event date： 2019.10.28 - 2019.10.31   Language：Japanese  

Event date： 2019.09.08 - 2019.09.11   Language：Japanese  

Event date： 2019.07.29 - 2019.08.01   Language：Japanese  

Event date： 2019.07.25 - 2019.07.28   Language：English  

Event date： 2019.03.14 - 2019.03.17   Language：Japanese  

Event date： 2018.11.28 - 2018.11.30   Language：Japanese  

Event date： 2018.11.23 - 2018.11.27   Language：Japanese  

Event date： 2018.11.04 - 2018.11.07   Language：Japanese  

Event date： 2018.07.09 - 2018.07.12   Language：English  

Event date： 2018.06.03 - 2018.06.08   Language：English  

Event date： 2018.03.01 - 2018.03.02   Language：Japanese  

Event date： 2017.11.08 - 2017.11.10   Language：Japanese  

Event date： 2017.09.03 - 2017.09.06   Language：Japanese  

Event date： 2017.07.20 - 2017.07.23   Language：English  

Event date： 2017.03.17 - 2017.03.20   Language：Japanese  

線虫C. elegansの神経系は302個の神経細胞から構成され，そのうち約170個が頭部に集中している．頭部に存在するこれら神経細胞の同時イメージングデータに対して，同期/非同期オーダーパラメータを用いた位相解析，および類似度行列に基づいたクラスタ解析を行った．本発表ではこれらの結果について報告する．また，シナプス結合に異常がある変異体での結果と野生型での結果の違いについても言及する．

Application no：特願2022-115074  Date applied：2022.07.19

Application no：2022-59045, ＰＣＴ／ＪＰ２０２３／０１０３３０  Date applied：2022.03.31

Application no：2022-035911  Date applied：2022.03.09

本発明の第１の観点は、サーマルカメラにより取得した映像データから脈拍を推定する非接触脈拍推定装置であって、前記映像データから抽出された複数の脈波信号に周波数解析を行う周波数解析部と、前記周波数解析部が出力した周波数成分のうち、脈拍の周波数である可能性がある周波数範囲である脈拍周波数範囲における最大の振幅を有する周波数成分を脈拍成分として決定する脈拍成分決定部とを備える、非接触脈拍推定装置である。

Application no：2021-189725  Date applied：2021.11.21

Presentation type：Invited lecture   Venue：名古屋大学  

Presentation type：Invited lecture   Venue：京都大学吉田キャンパス  

Presentation type：Invited lecture  

Presentation type：Invited lecture   Venue：大阪大学免疫学フロンティア研究センター  

Presentation type：Invited lecture  

Audience： Researchesrs,　General,　Company

Type：Other

Audience： Researchesrs,　General,　Company

Type：Seminar, workshop

Audience： Researchesrs,　General,　Company

Type：Lecture

北九州環境ミュージアムにおいて, 公益社団法人日本技術士会九州本部北九州地区支部2021年3月度CPDに講師として参加し, 「先端的な科学計測とデータサイエンスの理想的な協働へ向けて」という題目で講演を行った。

Audience： College students,　Graduate students

Type：Other

Type：Seminar, workshop