代码生成，python基于cnn和mfcc方法对CASIA的语音情感数据库的语音进行区分男女声音并进行情感分析

以下是基于CNN和MFCC方法对CASIA语音情感数据库进行性别识别和情感分析的Python代码示例：

首先，我们需要导入所需的库和数据集：

import os
import numpy as np
import pandas as pd
import librosa
import librosa.display
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten, Conv2D, MaxPooling2D
from tensorflow.keras.utils import to_categorical

# 设置数据集路径
data_path = "CASIA-3/CASIA-3/voice"
emotions = ["angry", "disgust", "fear", "happy", "neutral", "sad", "surprise"]

然后，我们可以定义一些辅助函数来处理数据。下面是一个函数，它将音频文件加载为MFCC（Mel频率倒谱系数）特征：

def extract_mfcc(file_path):
    audio, sr = librosa.load(file_path, res_type="kaiser_fast") # 加载音频文件
    mfccs = librosa.feature.mfcc(y=audio, sr=sr, n_mfcc=20) # 提取MFCC特征
    return mfccs

接下来，我们可以定义一个函数来加载数据集并将其转换为MFCC特征。我们还将为每个样本标签进行编码：

def load_data(test_size=0.2):
    X = []
    y = []
    for emotion in emotions:
        emotion_path = os.path.join(data_path, emotion)
        for file in os.listdir(emotion_path):
            file_path = os.path.join(emotion_path, file)
            mfccs = extract_mfcc(file_path)
            X.append(mfccs)
            y.append(emotion)
    X = np.array(X)
    y = np.array(y)
    le = LabelEncoder()
    y = to_categorical(le.fit_transform(y))
    return train_test_split(X, y, test_size=test_size, random_state=42)

现在，我们可以使用卷积神经网络（CNN）对MFCC特征进行分类。下面是一个简单的CNN模型：

def create_model(input_shape, num_classes):
    model = Sequential()
    model.add(Conv2D(32, (3, 3), padding='same', input_shape=input_shape))
    model.add(Activation('relu'))
    model.add(Conv2D(32, (3, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    model.add(Dropout(0.25))

    model.add(Conv2D(64, (3, 3), padding='same'))
    model.add(Activation('relu'))
    model.add(Conv2D(64, (3, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    model.add(Dropout(0.25))

    model.add(Flatten())
    model.add(Dense(512))
    model.add(Activation('relu'))
    model.add(Dropout(0.5))
    model.add(Dense(num_classes))
    model.add(Activation('softmax'))

    return model

最后，我们可以使用以下代码来训练模型并进行预测：

X_train, X_test, y_train, y_test = load_data() # 加载数据集

input_shape = (X_train.shape[1], X_train.shape[2], 1) # 定义输入形状
num_classes = y_train.shape[1] # 定义类别数

# 创建并编译模型
model = create_model(input_shape, num_classes)
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

# 训练模型
model.fit(X_train, y_train, batch_size=32, epochs=50, validation_data=(X_test, y_test))

# 预测新样本
new_sample = extract_mfcc("test_sample.wav")
new_sample = np.expand_dims(new_sample, axis=-1)
prediction = model.predict(new_sample)

希望这个代码示例能帮助你开始使用CNN和MFCC方法对语音情感数据库进行性别识别和情感分析！