RESUMO
In this paper, we propose a multi-scale mel domain feature map extraction algorithm to solve the problem that the speech recognition rate of dysarthria is difficult to improve. We used the empirical mode decomposition method to decompose speech signals and extracted Fbank features and their first-order differences for each of the three effective components to construct a new feature map, which could capture details in the frequency domain. Secondly, due to the problems of effective feature loss and high computational complexity in the training process of single channel neural network, we proposed a speech recognition network model in this paper. Finally, training and decoding were performed on the public UA-Speech dataset. The experimental results showed that the accuracy of the speech recognition model of this method reached 92.77%. Therefore, the algorithm proposed in this paper can effectively improve the speech recognition rate of dysarthria.
Assuntos
Humanos , Disartria/diagnóstico , Fala , Percepção da Fala , Algoritmos , Redes Neurais de ComputaçãoRESUMO
Interference of P with determination of Ca in food and the effect of interference suppressing agents of La, Sr and 8-hydroxyquinoline were studied. The results showed that 1) interference of P with Ca related to P/Ca ratio, the relative absorbance of Ca were getting increase along with, the decrease of P/Ca ratio if P/Ca ratio about 0.6; 2) La was more effective than the other interference suppressing agents. The relative absorbance of calcium reached to about 90% when 0.5% or 1% La was used in the P/Ca ratio of 30; 3) Sr had two actions. On one hand Sr had some suppressing action on P interference, and on the other hand Sr depressed the Ca absorption signal obviously; and 4) 8-hydroxyqui-noline had some but not very effective interference suppressing action. The content of Ca determined in food SRM were lower than certified value using 0.1% 8-hydroxyquinoline as interference suppressing agent.