基于改进混合蛙跳算法及SVM的耳语情感语音识别方法的DSP实现  

DSP Implementation of Whispered Speech Emotion Recognition Method Based on Improved Shuffled Frog Leaping Algorithm and SVM

作　　者：杨彦[1] 王浩[2] 赵力[2] 

机构地区：[1]盐城纺织职业技术学院,江苏盐城650500 [2]东南大学信息科学与工程学院,南京210096

出　　处：《电子器件》2012年第6期699-703,共5页Chinese Journal of Electron Devices

基　　金：国家自然科学基金项目(60872073;60975017;51075068);教育部博士点专项基金项目(20110092130004);江苏省高校自然科学基金项目(10KJB510005)

摘　　要：提出了一种基于改进混合蛙跳算法的SVM训练算法。该算法保持了混合蛙跳算法参数少和容易实现的特点,同时通过模拟退火的降温过程来提高算法的进化速度和精度。并用耳语情感语音识别实验来验证提出的基于改进混合蛙跳算法的SVM的有效性。实验结果表明,提出的新的模型的实验结果明显好于传统的SVM方法,证明了该方法的有效性。A training method of support vector machine （SVM）is proposed in this paper. lhe proposeu training method is based on an improved version of shuffled frog leaping algorithm （SFLA）. Two properties of SFLA, small number of parameters and easy realization, are retained in the proposed training method. Further, the evolution rate and precision of the proposed training method are improved through the temperature lowering process of the simulated annealing algorithm. The experiment of whispered speech emotion recognition is conducted to illustrate the effectiveness of the proposed training method. Experimental results reveal that the proposed training method performs obviously better than the traditional training method of SVM.

关 键 词：耳语情感语音 混合蛙跳算法 模拟退火 SVM 识别 

分 类 号：TN912.34[电子电信—通信与信息系统]