简体繁体 English

如何在STM32上计算梅尔谱图？

[英]How to compute the mel spectrogram on STM32?

原文 2022-10-22 16:35:59 8 1 c++/ neural-network/ embedded/ signal-processing/ stm32

I'm working with a STM32F4Discovery board, and I'm trying to run a neural.network on it for sound classification.我正在使用STM32F4Discovery板，我正在尝试在其上运行神经网络以进行声音分类。
I want to create the mel spectrogram directly on board from the sound recorded from the built-in microphone, which I already converted to PCM, and than give it as input to my neural.network (trained with tensorflow, and generated with STM Cube AI).我想直接从内置麦克风录制的声音创建梅尔频谱图，我已经将其转换为 PCM，然后将其作为我的神经网络的输入（使用 tensorflow 训练，并使用 STM Cube AI 生成).
The.network's input is an image of 30x30x1. The.network的输入是一张30x30x1的图片。
Is there a C/C++ library that can I use to implement it?是否有可用于实现它的 C/C++ 库？ I've tried LibrosaCpp but it crash when calling Eigen functions.我已经尝试过LibrosaCpp ，但它在调用 Eigen 函数时崩溃了。

1 个解决方案

It is possible to compute the mel spectrogram on STM32 boards by using CMSIS DSP Library provided by ARM alongside with the STM32_AI_AudioPreprocessing_Library included in the FP-AI-Sensing package .可以使用 ARM 提供的CMSIS DSP 库以及FP-AI-Sensing package中包含的STM32_AI_AudioPreprocessing_Library来计算 STM32 板上的梅尔频谱图。
In the FP-AI-Sensing library there is also an example for the mel spectrogram computation.在 FP-AI-Sensing 库中，还有一个梅尔频谱图计算示例。