欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对最小二乘支持向量机（LSSVM）在语音识别中识别速度慢的问题，提出了一种新的LSSVM稀疏化方法。采用独立成分分析（ICA）对语音特征降维，降低建模复杂度；将降维后的语音样本输入LSSVM建模；采用基于ICA的快速剪枝方法筛选支持向量，实现模型的稀疏化。韩语语音库的实验结果表明，在保持语音识别效果的基础上，语音识别速度明显提高。
英文摘要：
To solve the problem of slow recognition speed of least squares support vector machine （LSSVM） in speech recogni- tion, a novel LSSVM sparseness approach is proposed to achieve fast speech recognition. The method first utilizes independent component analysis （ICA） for dimension reduction of speech features, which decreases modeling complexity; then the dimension- reduced speech samples are inputted into LSSVM modeling; finally support vectors are screened with fast pruning technique based on ICA to achieve model sparseness, The experiments are carried out on the library of Korean phonetic database and the results indicate that the recognition speed is improved significantly while speech recognition effects are maintained.
同期刊论文项目
同项目期刊论文
Parameter optimization and application of support vector machine based on parallel artificial fish s
Study for classification of emotional speech by using optimized frame zero crossing with peak amplit
A robust feature extraction approach based on an auditory model for classification of speech and exp
期刊信息
位置：