http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
SPLICE 방법에 기반한 잡음 환경에서의 음성 인식 성능 향상
김종현,송화진,이종석,김형순,Kim, Jong-Hyeon,Song, Hwa-Jeon,Lee, Jong-Seok,Kim, Hyung-Soon 대한음성학회 2005 말소리 Vol.53 No.-
The performance of speech recognition system is degraded by mismatch between training and test environments. Recently, Stereo-based Piecewise LInear Compensation for Environments (SPLICE) was introduced to overcome environmental mismatch using stereo data. In this paper, we propose several methods to improve the conventional SPLICE and evaluate them in the Aurora2 task. We generalize SPLICE to compensate for covariance matrix as well as mean vector in the feature space, and thereby yielding the error rate reduction of 48.93%. We also employ the weighted sum of correction vectors using posterior probabilities of all Gaussians, and the error rate reduction of 48.62% is achieved. With the combination of the above two methods, the error rate is reduced by 49.61% from the Aurora2 baseline system.
김종현,구자예,오두석,Kim, Jong-Hyeon,Gu, Ja-Ye,O, Du-Suk 대한기계학회 1998 大韓機械學會論文集B Vol.22 No.5
Comparisons of joint probability density distribution obtained from the raw data of measured droplet sizes and velocities in a transient diesel fuel spray with computed joint probability density function were made. Simultaneous droplet sizes and velocities were obtained using PDPA. Mathematical probability density functions which can fit the experimental distributions were extracted using the principle of maximum likelihood. Through the statistical process of functions, mean droplet diameters, non-dimensional mass, momentum and kinetic energy were estimated and compared with the experimental ones. A joint log-hyperbolic density function presents quite well the experimental joint density distribution which were extracted from experimental data.