|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/audio-classification-using-convolutional-neural-net/
课程评论：没有评论
**酷课网（Coursera）课程总结：使用卷积神经网络进行音频分类** 本课程提供对处理音频文件进行机器学习的深入理解。您将学习如何使用 Python 从头到尾处理音频文件。课程重点介绍如何利用卷积神经网络（CNN）构建用于音频分类的 H5 AI 模型。 课程内容涵盖： * **音频文件处理：** 从音频环境识别、录音、切片（正负样本）、特征提取（时域、频域、声谱图）到数据预处理（裁剪、标记、批处理），为输入神经网络做好准备。 * **卷积神经网络（CNN）建模：** 学习如何使用 Python 构建和训练 H5 AI 模型，以实现音频预测。 * **树莓派 5 集成：** 掌握树莓派 5 的组装、编程、AI 模型部署以及在树莓派上进行音频文件预测。 * **项目实践：** 将训练好的 H5 AI 模型部署到树莓派 5 上，通过音频指令控制伺服电机运动，并进行实时音频预测测试。 本课程旨在帮助学习者全面掌握使用 CNN 进行音频分类的技术，并能将其应用于实际的嵌入式设备部署。（课程无详细教学大纲）
This course is designed to provide a real understanding of handling audio files in machine learning. This course will give you a complete track record of processing audio files from A to Z using Python. This course will explain how to use Convolutional Neural Networks to generate an H5 AI model for audio classification purposes. This course gives you a complete understanding of Raspberry Pi 5 assembly, programming, AI Model deployment, and prediction of audio files. We will learn how to identify audio environments for machine-learning purposes. We will learn how to record audio files and slice them into clips of positive and negative types. How to process the raw audio clips and inject the "keyword" to be detected by the neural network. Apply clip labeling, clip slicing, and clip batching for the preparation of feeding audio clips to the Neural Net. Apply the required stages (load, time domain, frequency domain, spectrogram, and resize) to process raw audio clips for prediction use. Use Python programming to generate an H5 AI model for audio prediction purposes. Deploy and run the H5 AI model inside the Raspberry Pi 5 to control the movement of the servo motor with audio order. Testing the model with a real-time audio prediction process.
