6个惊人的tensorflow js项目启动Web上的机器学习

As machine learning continues to accelerate, more and more languages are joining the bandwagon. JavaScript which has been the leader of the web ecosystem for a while now has slowly gained pace in machine learning as well.

随着机器学习的不断加速，越来越多的语言加入潮流。 JavaScript一直是网络生态系统的领导者，现在也慢慢地在机器学习方面取得了进步。

An encouraging number of JavaScript libraries, specifically for machine learning have released over the past year. This would certainly boost AI-powered web browsers. Additionally, JavaScript for machine learning has the advantage of quick and easy deployment of models in mobile applications by leveraging web views.

在过去的一年中，发布了数量惊人JavaScript库，专门用于机器学习。 这肯定会增强AI驱动的Web浏览器。 此外，用于机器学习JavaScript具有通过利用Web视图在移动应用程序中快速轻松地部署模型的优势。

At the heart of TensorFlow, we have a Web GL accelerated TensorFlow.js library that lets you train and run models directly in the browser or with Node.js. The best thing is, you don’t need to be a machine learning expert in order to deploy basic models as the library already provides a bunch of pre-trained models for classification, segmentation, and more.

在TensorFlow的核心，我们有一个Web GL加速的TensorFlow.js库，可让您直接在浏览器中或使用Node.js训练和运行模型。 最好的事情是，您不需要成为机器学习专家即可部署基本模型，因为该库已经提供了一堆经过预训练的分类，细分等模型。

Let’s look at a few awesome libraries build on top of TensorFlow.js to inspire your next machine learning project.

让我们看一下在TensorFlow.js之上构建的一些很棒的库，以激发您的下一个机器学习项目。

1.面部和情绪识别 (1. Face And Emotion Recognition)

Face detection has been one of the classic use cases in Open-CV and naturally, it became the common use case in machine learning as well.

人脸检测已成为Open-CV中的经典用例之一，自然也成为了机器学习中的常见用例。

face-api.js is a JavaScript face recognition library implemented on top of TensorFlow.js. It lets you detect and recognize faces and landmarks while also determining emotions and gender in images.

face-api.js是在TensorFlow.js之上实现JavaScript人脸识别库。 它使您可以检测和识别面部和地标，同时还可以确定图像中的情绪和性别。

The library completely abstracts away the underlying implementation to provide you with an easy to use high-level API. All you need to do is invoke the methods for the relevant neural net model with an option to also call the ready-drawing functions for overlaying the feature points on the canvas.

该库完全抽象出了基础实现，为您提供了易于使用的高级API。 您所需要做的就是调用相关神经网络模型的方法，并带有一个选项，该选项还调用现成的绘制函数以将特征点覆盖在画布上。