著者

Mariana RODRIGUES MAKIUCHI Tifani WARNITA Nakamasa INOUE Koichi SHINODA Michitaka YOSHIMURA Momoko KITAZAWA Kei FUNAKI Yoko EGUCHI Taishiro KISHIMOTO

出版者

The Institute of Electronics, Information and Communication Engineers

雑誌

IEICE Transactions on Information and Systems (ISSN:09168532)

巻号頁・発行日

vol.E104.D, no.11, pp.1930-1940, 2021-11-01 (Released:2021-11-01)

参考文献数

71

被引用文献数

6

---

We propose a non-invasive and cost-effective method to automatically detect dementia by utilizing solely speech audio data. We extract paralinguistic features for a short speech segment and use Gated Convolutional Neural Networks (GCNN) to classify it into dementia or healthy. We evaluate our method on the Pitt Corpus and on our own dataset, the PROMPT Database. Our method yields the accuracy of 73.1% on the Pitt Corpus using an average of 114 seconds of speech data. In the PROMPT Database, our method yields the accuracy of 74.7% using 4 seconds of speech data and it improves to 80.8% when we use all the patient's speech data. Furthermore, we evaluate our method on a three-class classification problem in which we included the Mild Cognitive Impairment (MCI) class and achieved the accuracy of 60.6% with 40 seconds of speech data.

著者

Ayako Sekine Takatomo Watanabe Genki Naruse Ayae Takada Shingo Fujimoto Noriko Ozawa Koichi Shinoda Yuzuru Nohisa Hiroyuki Okura

出版者

The Japanese Circulation Society

雑誌

Circulation Journal (ISSN:13469843)

巻号頁・発行日

vol.85, no.1, pp.78, 2020-12-25 (Released:2020-12-25)

参考文献数

1

著者

Shiro Mori Takuya Yamane Masaki Yahata Koichi Shinoda Naho Murata

出版者

一般社団法人 園芸学会

雑誌

The Horticulture Journal (ISSN:21890102)

巻号頁・発行日

pp.MI-117, (Released:2016-03-25)

被引用文献数

9

---

Limonium bellidifolium (2n = 2x = 18), a perennial statice belonging to the family Plumbaginaceae, is cultivated as a garden plant or for cut flowers and is an important breeding material for the production of hybrid cultivars in the genus. In this study, chromosome doubling in L. bellidifolium was attempted to increase the variability in horticultural traits. Seeds of this species were treated with an antimitotic agent, colchicine, at different concentrations and exposure periods. The treated seeds were sown in soil in a cell tray and the seedlings were grown in a greenhouse. More than 50% of the seedlings treated with colchicine for 24 or 48 h, irrespective of the concentration, survived for 4 months after treatment. Most of the seedlings treated for 72 h at any concentration showed very poor growth and abnormal thickening of the hypocotyl, and ultimately died. The surviving seedlings were grown in 9-cm pots. Flow cytometry analysis of leaf tissues showed that 2.5%–5% of plants that received 0.05% colchicine for 72 h, 0.25% colchicine for 24 h, 0.25% colchicine for 48 h, or 0.5% colchicine for 48 h were tetraploid (4x) or mixoploid (2x + 4x). The highest frequencies of tetraploids and mixoploids occurred following treatment with 0.05% colchicine for 72 h. These results showed that colchicine treatment of seeds is effective for chromosome doubling in L. bellidifolium. After 3 years of cultivation, the morphological characteristics of diploid and tetraploid L. bellidifolium plants at the flowering stage were investigated. The stomatal density was lower in all investigated tetraploid and mixoploid plants than in the control diploid plant. The stomatal length was 1.1- to 1.5-fold higher in all tetraploid and mixoploid plants than in the control. Tetraploid plants tended to have wider and thicker leaves than the control and also produced larger flowers.

著者

篠田 浩一 Koichi Shinoda 東京工業大学大学院情報理工学研究科 School of Information Science and Technology Tokyo Institute of Technology

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.27, no.4, pp.359-364, 2012-07-01

著者

Takafumi KOSHINAKA Kentaro NAGATOMO Koichi SHINODA

出版者

The Institute of Electronics, Information and Communication Engineers

雑誌

IEICE TRANSACTIONS on Information and Systems (ISSN:09168532)

巻号頁・発行日

vol.E95-D, no.10, pp.2469-2478, 2012-10-01

---

A novel online speaker clustering method based on a generative model is proposed. It employs an incremental variant of variational Bayesian learning and provides probabilistic (non-deterministic) decisions for each input utterance, on the basis of the history of preceding utterances. It can be expected to be robust against errors in cluster estimation and the classification of utterances, and hence to be applicable to many real-time applications. Experimental results show that it produces 50% fewer classification errors than does a conventional online method. They also show that it is possible to reduce the number of speech recognition errors by combining the method with unsupervised speaker adaptation.