A5086456461- Face and Expression Recognition
- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Anomaly Detection Techniques and Applications
- Neural Networks and Applications
- Face recognition and analysis
- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Image Processing Techniques and Applications
- Medical Image Segmentation Techniques
- Human Pose and Action Recognition
- Spectroscopy and Chemometric Analyses
- Robotics and Sensor-Based Localization
- Domain Adaptation and Few-Shot Learning
- Advanced Image Processing Techniques
- Remote-Sensing Image Classification
- Machine Learning and Data Classification
- Blind Source Separation Techniques
- Image and Signal Denoising Methods
- Image and Object Detection Techniques
- Advanced Algorithms and Applications
- Cell Image Analysis Techniques
- Machine Learning and ELM
- Generative Adversarial Networks and Image Synthesis
Hiroshima University
2016-2025
Higashihiroshima Medical Center
2017-2023
Hiroshima University of Economics
2011-2022
Tokyo Denki University
2012-2014
National Institute of Advanced Industrial Science and Technology
2003-2013
University of Tsukuba
2008-2009
Fujitsu (Japan)
2005
Chukyo University
2004
Kwansei Gakuin University
2004
Gifu University
2004
This paper introduces the sparse regularization for convolutional neural network (CNN) with rectified linear units (ReLU) in hidden layers. By introducing sparseness inputs of ReLU, there is effect to push ReLU zero learning process. Thus it expected that unnecessary increase outputs can be prevented. similar Batch Normalization. Also negative values reduced by sparseness. improve generalization trained network. The relations between proposed approach and Normalization or modifications...
Abstract Background Early diagnosis of osteoporosis can potentially decrease the risk fractures and improve quality life. Detection thin inferior cortices mandible on dental panoramic radiographs could be useful for identifying postmenopausal women with low bone mineral density (BMD) or osteoporosis. The aim our study was to assess diagnostic efficacy using kernel-based support vector machine (SVM) learning regarding cortical width identify BMD. Methods We employed newly adopted SVM method...
Gives a basic idea and its fundamental algorithms of the visual interface for image database systems. The QVE (Query by Visual Example) accepts sketch roughly drawn user to retrieve original similar images. system evaluates similarity between rough sketch, i.e. example, each data in automatically. is implemented examined on an experimental electronic art gallery called ART MUSEUM. This paper also gives some results current evaluation. are quite effective content based retrieval.< <ETX...
This paper proposes a crowd counting method. Crowd is difficult because of large appearance changes target which caused by density and scale changes. Conventional methods generally utilize one predictor (e,g., regression multi-class classifier). However, such only can not count targets with well. In this paper, we propose to predict the number using multiple CNNs specialized specific appearance, those are adaptively selected according test image. By integrating CNNs, proposed method has...
Proposes a face recognition method which is characterized by structural simplicity, trainability and high speed. The consists of two stages feature extractions: first, higher order local autocorrelation features are shift-invariant additive extracted from an input image; then those linearly combined on the basis multivariate analysis methods so as to provide new effective for in learning examples.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Describes how the authors have combined speech recognition, dialogue management, and statistical learning procedures to develop Jijo-2; an office robot that can communicate with humans learn about its environment.
Pluripotent stem cells can potentially be used in clinical applications as a model for studying disease progress. This tracking of disease-causing events requires constant assessment the quality cells. Existing approaches are inadequate robust and automated differentiation cell colonies. In this study, we developed new vector–based convolutional neural network (V-CNN) with respect to extracted features induced pluripotent (iPSC) colony distinguishing characteristics. A transfer function from...
A complex autoregressive model for invariant feature extraction to recognize arbitrary shapes on a plane is presented. fast algorithm calculate coefficients and PARCOR of the also shown. The are rotation around origin choice starting point in tracing boundary. It possible make them scale translation. Experimental results that complicated like nonconvex boundaries can be recognized high accuracy, even low-order model. seen tend provide more accurate classification than AR coefficients.< <ETX...
This study proposed a new automated screening system based on hybrid genetic swarm fuzzy (GSF) classifier using digital dental panoramic radiographs to diagnose females with low bone mineral density (BMD) or osteoporosis.The geometrical attributes of both the mandibular cortical and trabecular were acquired previously developed software. Designing an for osteoporosis involved partitioning input generate initial membership function (MF) rule set (RS), classification inference optimization...
This paper proposes a scale invariant face detection method which combines higher-order local autocorrelation (HLAC) features extracted from log-polar transformed image with linear discriminant analysis for "face" and "not face" classification. Since HLAC of images are sensitive to shifts face, we utilize this property develop method. become rotation because scalings rotations expressed as in (coordinate). By combining these the is extended treat classes, system can be realized.
Describes how the authors have combined speech recognition, dialogue management, and statistical learning procedures to develop Jijo-2; an office robot that can communicate with humans learn about its environment.
In this paper, we develop a visualization tool suitable for deep neural networks (DNN). Although typical dimensionality reduction methods, such as principal component analysis (PCA), are useful to visualize highdimensional data 2 or 3 dimensional representations, most of those methods focus their attention on how create essential subspaces based only given unique feature representation. On the other hand, DNN naturally have consecutive multiple representations corresponding intermediate...
We have built a prediction model of the fluorescence quantum yields metalloles. Based on suggestion by model, we synthesized 10 fluorescent molecules to confirm accuracy. By measuring molecules, it was demonstrated that our reasonably classified with an accuracy 0.7. In particular, low were perfectly predicted for demonstrating usefulness screen out weakly from candidates. On other hand, precision 0.5 attributed bias in training dataset containing many fluorine-containing high yields. Our...