针对语音识别任务——10个数字的孤立词识别，分别使用GMM 和 HMM 建立对 0～9 十个数字的孤立词语音分类识别模型的代码python语言

下面是使用GMM和HMM建立对10个数字的孤立词语音分类识别模型的Python代码示例：

1. 使用GMM进行特征建模和分类

import numpy as np
import librosa
from sklearn.mixture import GaussianMixture

# 加载语音数据，提取MFCC特征
def extract_features(file_name):
    y, sr = librosa.load(file_name)
    mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=20)
    return np.mean(mfccs, axis=1)

# 训练GMM模型
def train_gmm_model(features):
    gmm = GaussianMixture(n_components=10, max_iter=200, covariance_type='diag', n_init=3)
    gmm.fit(features)
    return gmm

# 识别测试语音
def recognize_gmm(gmm_models, test_feature):
    scores = []
    for i in range(10):
        score = gmm_models[i].score(test_feature.reshape(1, -1))
        scores.append(score)
    return np.argmax(scores)

# 加载训练数据，训练GMM模型
def train_gmm():
    gmm_models = []
    for i in range(10):
        file_name = 'digit_' + str(i) + '.wav'
        features = extract_features(file_name)
        gmm = train_gmm_model(features)
        gmm_models.append(gmm)
    return gmm_models

# 测试模型
def test_gmm(gmm_models):
    correct = 0
    total = 0
    for i in range(10):
        for j in range(10):
            file_name = 'digit_' + str(i) + '_' + str(j) + '.wav'
            features = extract_features(file_name)
            label = recognize_gmm(gmm_models, features)
            if label == i:
                correct += 1
            total += 1
    print('Accuracy:', correct / total)

# 训练和测试GMM模型
gmm_models = train_gmm()
test_gmm(gmm_models)

2. 使用HMM进行分类

import numpy as np
import librosa
from hmmlearn import hmm

# 加载语音数据，提取MFCC特征
def extract_features(file_name):
    y, sr = librosa.load(file_name)
    mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=20)
    return mfccs.T

# 训练HMM模型
def train_hmm_model(features):
    model = hmm.GaussianHMM(n_components=5, covariance_type='diag', n_iter=100)
    model.fit(features)
    return model

# 识别测试语音
def recognize_hmm(hmm_models, test_feature):
    scores = []
    for i in range(10):
        score = hmm_models[i].score(test_feature)
        scores.append(score)
    return np.argmax(scores)

# 加载训练数据，训练HMM模型
def train_hmm():
    hmm_models = []
    for i in range(10):
        file_name = 'digit_' + str(i) + '.wav'
        features = extract_features(file_name)
        model = train_hmm_model(features)
        hmm_models.append(model)
    return hmm_models

# 测试模型
def test_hmm(hmm_models):
    correct = 0
    total = 0
    for i in range(10):
        for j in range(10):
            file_name = 'digit_' + str(i) + '_' + str(j) + '.wav'
            features = extract_features(file_name)
            label = recognize_hmm(hmm_models, features)
            if label == i:
                correct += 1
            total += 1
    print('Accuracy:', correct / total)

# 训练和测试HMM模型
hmm_models = train_hmm()
test_hmm(hmm_models)

以上是使用GMM和HMM进行10个数字的孤立词识别的Python代码示例，其中使用了librosa和sklearn库进行特征提取和GMM建模，使用了hmmlearn库进行HMM建模。