Search Results 1-20 of 107

  • Deep Metric Learning with Triplet-Margin-Center Loss for Sketch Face Recognition

    FENG Yujian , WU Fei , JI Yimu , JING Xiao-Yuan , YU Jian

    … In this paper, we propose triplet-margin-center loss to cope with the above problem by combining the triplet loss and center loss. … The triplet-margin-center loss can enlarge the distance of inter-class samples and reduce intra-class sample variations simultaneously, and improve intra-class compactness. … Moreover, the triplet-margin-center loss applies a hard triplet sample selection strategy. …

    IEICE Transactions on Information and Systems E103.D(11), 2394-2397, 2020

    J-STAGE 

  • Automatic Construction of a Large-Scale Speech Recognition Database Using Multi-Genre Broadcast Data with Inaccurate Subtitle Timestamps

    BANG Jeong-Uk , CHOI Mu-Yeol , KIM Sang-Hun , KWON Oh-Wook

    … <p>As deep learning-based speech recognition systems are spotlighted, the need for large-scale speech databases for acoustic model training is increasing. … The proposed method first extracts the most likely speech intervals by removing subtitle texts with low subtitle quality index, concatenating adjacent subtitle texts into a merged subtitle text, and adding a margin to the timestamp of the merged subtitle text. …

    IEICE Transactions on Information and Systems E103.D(2), 406-415, 2020

    J-STAGE 

  • STUDY ON MEASUREMENT AND EVALUATION OF FLYING QUALITY OF LIGHT AIRCRAFT  [in Japanese]

    浅見 将輝

    … For small aircraft design, good flying quality is important to keep safety margin and to learn how to fly. … Recently new category of small aircraft so called "Light Sport Aircraft: LSA" is spreading worldwide, the LSA is one of the categories of potential popular training aircraft. …

    法政大学大学院紀要. 理工学・工学研究科編 (60), 1-8, 2019-03-31

    IR  DOI 

  • Artificial Neural Network-Based QoT Estimation for Lightpath Provisioning in Optical Networks

    ZHANG Min , XU Bo , LI Xiaoyun , FU Dong , LIU Jian , WU Baojian , QIU Kun

    … Though it is common to use worst case assumptions for estimating the quality of transmission (QoT) in the physical layer, over provisioning results in high margin requirements. … High estimation accuracy of over 98% with estimation errors of less than 0.5dB can be achieved given enough training data. …

    IEICE Transactions on Communications E102.B(11), 2104-2112, 2019

    J-STAGE 

  • Super-Resolution Simulation for Real-Time Prediction of Urban Micrometeorology

    Onishi Ryo , Sugiyama Daisuke , Matsuda Keigo

    … We performed HR micrometeorology simulations on a supercomputer to obtain datasets for training the CNN in the SR method. … It is shown that the proposed SR method can be used with a spatial scaling factor of 4 and that it outperforms conventional interpolation methods by a large margin. …

    SOLA 15(0), 178-182, 2019

    J-STAGE 

  • Academic and Vocational aspects of the Field Classification of Tertiary Education in Japan  [in Japanese]

    吉本 圭一

    … Having gone through these considerations, the importance of human resources development and technical and vocational education was confirmed and the possibilities of integrated education and training with international recognition and the direction of quality assurance were examined. … As the results, the classifications of academics, industries and occupations and education and training were inter-correlating. …

    大学院教育学研究紀要 (21), 25-44, 2019

    IR 

  • Effect of Equality Constraints to Unconstrained Large Margin Distribution Machines

    Abe Shigeo

    … Abstract Unconstrained large margin distribution machines (ULDMs) maximize the margin mean and minimize the margin variance without constraints. … By setting a hyperparameter to control the trade-off between the generalization ability and the training error to zero, LS LDMs reduce to ULDMs. …

    Lecture Notes in Computer Science (11081), 41-53, 2018-08-30

    IR 

  • Development and Experimental Ealuation of Large Geometric Margin Minimum Classification Error Training for Classification of Variable-length Time Series Patterns  [in Japanese]

    松廣 達也 , 橋本 哲也 , 北岡 見生代 , ア デイビッド , 落合 翼 , 渡辺 秀行 , 片桐 滋 , 大崎 美穂

    … Recently, one of the latest discriminative training methods, Minimum Classification Error (MCE) training, was extended to Large Geometric Margin MCE (LGM-MCE) training. …

    情報処理学会論文誌 59(4), 1295-1308, 2018-04-15

    IPSJ 

  • Oncoplastic Breast-conserving Surgery - Bridging the Gap between Mastectomy and Breast-conserving Surgery  [in Japanese]

    Fujimoto Hiroshi , Nagashima Takeshi , Ohtsuka Masayuki

    … To reduce the positive margin, it is necessary to select appropriate candidates based on the careful evaluation of breast images and tumor characteristics. … Expertise in both oncology and plastic surgery is necessary for success, and the training system needs to be improved.</p> …

    Oncoplastic Breast Surgery 3(3-4), 59-68, 2018

    J-STAGE  Ichushi Web 

  • Unconstrained large margin distribution machines

    Abe Shigeo

    … Large margin distribution machines (LDMs) maximize the margin mean and minimize the margin variance, and show good generalization performance compared to support vector machines (SVMs). … In this paper we propose unconstrained large margin distribution machines (ULDMs). …

    Pattern Recognition Letters (98), 96-102, 2017-10

    IR 

  • Marketing and value chain system of marine frozen fish in Bangladesh

    Islam M. Serajul , Haque Md. Mojammel , Hasan Md. Kamrul

    … High priced fish demanded a high marketing cost resulting higher marketing margin and profit compared to low priced fish. … Actually, marketing margin of respective species of marine frozen fish at each level of market is the added value of that particular species of fish.</p> …

    Bulletin of the Institute of Tropical Agriculture, Kyushu University 40(0), 63-75, 2017

    J-STAGE 

  • On Information Requirement for a Pilot's Altitude Control using Elevator.(For more Effective Bridge Training)  [in Japanese]

    上村 常治 , エントジンガー ヨルグ オノ , 森 亮太 , 松永 大一郎 , 鈴木 真二

    … However, recently I realized the lack of 'Phase Margin' between elevator and altitude control, from a feedback control theoretical analysis. … This knowledge will help flight training, especially 'Bridge Training' which Pilots receive when transitioning from small to big aircraft.JSASS-2017-5237 …

    飛行機シンポジウム講演集 (55), 2017

    IR 

  • LGM-MCE Training Based on Parzen Estimation with Sample Weighting  [in Japanese]

    2016年度 情報処理学会関西支部 支部大会 講演論文集 (2016), 2016-09-16

    IPSJ 

  • A Study on Improvements to the Margin of Error in Personnel Evaluation : Focus on Perspective on the Actual Operation of the Personnel Evaluation System  [in Japanese]

    INOUE Hitoshi

    … The results gained are applied to the decision making process concerning an employee's promotion, salary increases, bonuses, transfers, and further job training. …

    Osaka Sangyo University journal of business administration 17(1・2), 31-51, 2016-02

  • Experimental Evaluation of Speech Recognition using Large Geometric Margin Minimum Classification Error Training  [in Japanese]

    松廣 達也 , 北岡 見生代 , ア デイビッド , 渡辺 秀行 , 片桐 滋 , 大崎 美穂

    情報処理学会関西支部支部大会講演論文集, 6p, 2016

  • The Development of a Home-use Method for Evaluating Muscular Strength of Lower Limbs  [in Japanese]

    YOSHIOKA Shinsuke , NAGANO Akinori

    … However, according to the suggestions in previous studies, we can prevent the muscular strength from decreasing by performing appropriate strength training in daily life. … To advance the training appropriately, it is necessary to evaluate current muscular strength objectively and to ascertain effectiveness of the training. … Therefore, we define a new index of muscular strength of the lower limbs, which we named “muscular strength margin”. …

    Transactions of Japanese Society for Medical and Biological Engineering 54(3), 112-119, 2016

    J-STAGE  Ichushi Web 

  • New Softness Evaluation Tissue Softness Analyzer TSA  [in Japanese]

    Miyaoka Hiroyuki , Sasaki Kunio

    … High exactness requirements present very high demands concerning the selection and the training of the members of the panel as well as the mathematic and statistics analysis. … Furthermore the testing results of a panel are dependent on testing procedure and the specifics of the area (cultural preferences, climate…) But even if these requirements are fulfilled at best, it is definitely not possible to undercut a certain error margin. …

    JAPAN TAPPI JOURNAL 70(5), 505-508, 2016

    J-STAGE 

  • Protein Fold Classification Using Large Margin Combination of Distance Metrics

    SURYANTO Chendra Hadi , FUKUI Kazuhiro , HINO Hideitsu

    … In this paper, by generalizing the concept of the large margin nearest neighbor (LMNN), a method for combining multiple distance metrics from different types of protein structure comparison methods for protein fold classification task is proposed. …

    IEICE Transactions on Information and Systems E99.D(3), 714-723, 2016

    J-STAGE 

  • Evaluation of Automatic Prototype-Model Size Optimization in Large Geometric Margin Minimum Classification Error Training  [in Japanese]

    荻野 将大 , 渡辺 秀行 , 片桐 滋 [他]

    電子情報通信学会技術研究報告 = IEICE technical report : 信学技報 115(388), 1-6, 2015-12-21

  • Relation between Data Grouping and Robustness to Unseen Data in Large Geometric Margin Minimum Classification Error Training  [in Japanese]

    SHIRAISHI Hiroyuki , WATANABE Hideyuki , KATAGIRI Shigeru , Lu XUGANG , HORI Chiori , OHSAKI Miho

    … To develop a pattern classifier that is robust to unseen pattern samples, classifier parameters have been conventionally trained using both training and validation sample sets. … However, there are no clear criteria for dividing the samples in hand into training and validation sets. … In addition, such grouping decreases the number of samples for both training and validation, often lowering robustness to unseen samples. …

    Technical report of IEICE. Multimedia and virtual environment 114(410), 177-182, 2015-01-22

Page Top