当前位置:网站首页>Voiceprint Technology (IV): Engineering deployment of voiceprint recognition
Voiceprint Technology (IV): Engineering deployment of voiceprint recognition
2022-06-25 09:05:00 【u013250861】
4.1 From model to product
4.1.1 Model is not equal to product
For academia , Through training data , A group of voiceprint recognition models are trained by a specific algorithm , And verify its performance on the test data —— These jobs add up , Enough to constitute a complete research project . If there are some innovations or changes in data processing or training algorithms , And add some other methods to compare the experimental results , Then this work is enough to write a research paper with rich content , To publish in an academic conference or journal .
however , For the industry , Have a voice print recognition model with good performance , It's just the first step in the long march . It is necessary to deploy voiceprint recognition technology to practical and easy-to-use products , There are countless engineering problems to be solved , For example, the architecture of the system 、 Model release 、 Software maintenance , And the interaction between users and products , wait . This chapter will introduce some of the most common engineering problems , The basic architectures of the three deployment schemes are briefly described .
4.1.2 Basic concepts of software engineering
Before introducing the engineering problems in voiceprint recognition , Considering that some readers of this book may not have any actual product level software development experience , The author first gives a very brief introduction to some basic concepts in software engineering . Although these concepts may have little to do with voiceprint recognition itself , But for readers who intend to enter the enterprise for actual product development in the future , These concepts are very important , It may even determine the success or failure of a project or a product .
Software engineering itself as a discipline , It covers a wide range of contents , Therefore, there are a vast number of books and tutorials on software engineering . For readers with limited time , Consider briefly reading the most famous 《 The way of programmer training : The Pragmatic Programmer 》 A Book [106], Or read Google Voice engineers directly Fergus Henderson The author's thesis Software Engineering at Google [107].
1. The scale and life cycle of software
For academic research or publishing papers ,
边栏推荐
- sklearn PolynomialFeatures的具体用法
- compiling stm32f4xx_ it. c... “.\Objects\BH-F407.axf“ - 42 Error(s), 1 Warning(s).
- 自定义注解之编译时注解(RetentionPolicy.CLASS)
- Notes on key words in the original English work biography of jobs (IV) [chapter two]
- Jmeter接口测试,关联接口实现步骤(token)
- Stimulsoft ultimate presents reports and dashboards
- QSS buttons of different styles
- C # startup program loses double quotation marks for parameters passed. How to solve it?
- Specific usage of sklearn polynomialfeatures
- matplotlib matplotlib中plt.grid()
猜你喜欢
A 35 year old Tencent employee was laid off and sighed: a suite in Beijing, with a deposit of more than 7 million, was anxious about unemployment
matplotlib matplotlib中axvline()和axhline()函数
Oracle-单行函数大全
sklearn 高维数据集制作make_circles 和 make_moons
How to become a software testing expert? From 3K to 17k a month, what have I done?
高速缓冲存储器Cache的映射方式
compiling stm32f4xx_it.c... “.\Objects\BH-F407.axf“ - 42 Error(s), 1 Warning(s).
五、项目实战---识别人和马
【无标题】**数据库课设:三天完成学生信息管理系统**
Unity--configurable joint -- a simple tutorial to get you started with configurable joints
随机推荐
matplotlib matplotlib中决策边界绘制函数plot_decision_boundary和plt.contourf函数详解
matplotlib matplotlib中plt.grid()
Analysis of a video website m3u8 non perceptual encryption
微服务调用组件Ribbon底层调用流程分析
annotation lib 和 processor lib如何引用
2、 Training fashion_ MNIST dataset
声纹技术(三):声纹识别技术
Notes on key words in the original English work biography of jobs (IV) [chapter two]
Unity--Configurable Joint——简单教程,带你入门可配置关节
sklearn 高维数据集制作make_circles 和 make_moons
How to solve the 10061 error of MySQL in Linux
Explanation of assertions in JMeter
声纹技术(七):声纹技术的未来
第十五周作业
socket编程——poll模型
IC研发常用英文术语缩写
Analysis on the bottom calling process of micro service calling component ribbon
Jmeter中的断言使用讲解
二、训练fashion_mnist数据集
Object. Can defineproperty also listen for array changes?