请问你可以提供一份CNN识别KTH源代码吗

当然可以，下面是使用Python和TensorFlow框架实现的CNN识别KTH数据集的源代码：

import tensorflow as tf
import numpy as np
import os
import cv2

# 数据路径
data_dir = './data/KTH_Action_Dataset/'
train_dir = os.path.join(data_dir, 'train/')
test_dir = os.path.join(data_dir, 'test/')

# 训练集和测试集的大小
train_size = 2400
test_size = 600

# 定义标签
label_names = ['boxing', 'handclapping', 'handwaving', 'jogging', 'running', 'walking']
num_classes = len(label_names)
label_to_index = dict(zip(label_names, range(num_classes)))

# 加载数据函数
def load_data(data_dir, num_samples):
    images = []
    labels = []
    for label_name in label_names:
        dir_name = os.path.join(data_dir, label_name)
        class_index = label_to_index[label_name]
        for i, file_name in enumerate(os.listdir(dir_name)):
            if i >= num_samples // num_classes:
                break
            # 加载图像并将其调整为统一大小
            image = cv2.imread(os.path.join(dir_name, file_name))
            image = cv2.resize(image, (80, 80))
            images.append(image)
            labels.append(class_index)
    return np.array(images), np.array(labels)

# 加载训练集和测试集
train_images, train_labels = load_data(train_dir, train_size)
test_images, test_labels = load_data(test_dir, test_size)

# 数据预处理
train_images = train_images.astype('float32') / 255
test_images = test_images.astype('float32') / 255

# 将标签转换为独热编码
train_labels = tf.keras.utils.to_categorical(train_labels, num_classes)
test_labels = tf.keras.utils.to_categorical(test_labels, num_classes)

# 定义模型
model = tf.keras.Sequential([
    tf.keras.layers.Conv2D(32, (3, 3), activation='relu', input_shape=(80, 80, 3)),
    tf.keras.layers.MaxPooling2D((2, 2)),
    tf.keras.layers.Conv2D(64, (3, 3), activation='relu'),
    tf.keras.layers.MaxPooling2D((2, 2)),
    tf.keras.layers.Conv2D(128, (3, 3), activation='relu'),
    tf.keras.layers.MaxPooling2D((2, 2)),
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dense(num_classes, activation='softmax')
])

# 编译模型
model.compile(optimizer='adam',
              loss='categorical_crossentropy',
              metrics=['accuracy'])

# 训练模型
model.fit(train_images, train_labels, epochs=10, batch_size=32)

# 评估模型
test_loss, test_acc = model.evaluate(test_images, test_labels, verbose=2)
print('Test accuracy:', test_acc)

在这个示例代码中，我们使用了KTH数据集作为示例数据，通过加载数据、预处理数据、定义模型、编译模型、训练模型和评估模型等步骤，实现了一个简单的CNN模型来识别KTH数据集中的动作类别。你可以根据自己的需求修改代码并应用到自己的项目中。