您现在访问的是微软AZURE全球版技术文档网站,若需要访问由世纪互联运营的MICROSOFT AZURE中国区技术文档网站,请访问 https://docs.azure.cn.

什么是自定义视觉?What is Custom Vision?

重要

现在,将对此服务的所有 HTTP 请求强制执行 TLS 1.2。TLS 1.2 is now enforced for all HTTP requests to this service. 有关详细信息,请参阅 Azure 认知服务安全性For more information, see Azure Cognitive Services security.

Azure 自定义视觉是一种图像识别服务,可用于构建、部署和改进你自己的图像标识符。Azure Custom Vision is an image recognition service that lets you build, deploy, and improve your own image identifiers. 图像标识符根据图像的视觉特征将标签(表示类或对象)应用于图像。An image identifier applies labels (which represent classes or objects) to images, according to their visual characteristics. 计算机视觉服务不同,自定义视觉允许你指定标签并训练自定义模型以对其进行检测。Unlike the Computer Vision service, Custom Vision allows you to specify the labels and train custom models to detect them.

作用What it does

自定义视觉服务使用机器学习算法分析图像。The Custom Vision service uses a machine learning algorithm to analyze images. 你作为开发人员提交多组图像,其中包含或缺少相关的特征。You, the developer, submit groups of images that feature and lack the characteristics in question. 请在提交时自行标记图像。You label the images yourself at the time of submission. 然后,此算法会针对该数据进行训练并计算其自己的准确度,方法是针对这些相同的图像自行进行测试。Then, the algorithm trains to this data and calculates its own accuracy by testing itself on those same images. 训练算法以后,即可在图像识别应用中测试它、重新训练它并最终使用它对新图像进行分类。Once the algorithm is trained, you can test, retrain, and eventually use it in your image recognition app to classify new images. 也可导出模型本身,方便脱机使用。You can also export the model itself for offline use.

分类和对象检测Classification and object detection

可以将自定义视觉功能分为两种功能。Custom Vision functionality can be divided into two features. 图像分类 可将一个或多个标签应用到图像。Image classification applies one or more labels to an image. 对象检测 与之类似,但还在图像中返回坐标,坐标中可以找到应用的标签。Object detection is similar, but it also returns the coordinates in the image where the applied label(s) can be found.

OptimizationOptimization

自定义视觉服务经过优化,可以快速识别图像之间的主要差异,因此你可以使用少量数据开始原型制作。The Custom Vision service is optimized to quickly recognize major differences between images, so you can start prototyping your model with a small amount of data. 开始时,每个标签通常可以包含 50 个图像。50 images per label are generally a good start. 此服务不适用于检测图像中的细微差异(例如,在质量保证方案中检测细微裂纹或凹陷)。However, the service is not optimal for detecting subtle differences in images (for example, detecting minor cracks or dents in quality assurance scenarios).

另外,可以从多种自定义视觉算法中进行选择,这些算法已针对包含某些主题内容(例如特征点或零售商品)的图像进行了优化。Additionally, you can choose from several varieties of the Custom Vision algorithm that are optimized for images with certain subject material—for example, landmarks or retail items. 有关详细信息,请参阅生成分类器生成对象检测器指南。For more information, see the Build a classifier or Build an object detector guides.

组成部分What it includes

自定义视觉服务以一组本机 SDK 的形式提供,另外也在自定义视觉网站的基于 Web 的界面上提供。The Custom Vision Service is available as a set of native SDKs as well as through a web-based interface on the Custom Vision website. 可以通过一种或两种界面来创建、测试和训练模型。You can create, test, and train a model through either interface or use both together.

Chrome 浏览器窗口中的自定义视觉网站

数据隐私和安全性Data privacy and security

与所有认知服务一样,使用自定义视觉服务的开发人员应该了解 Microsoft 针对客户数据的政策。As with all of the Cognitive Services, developers using the Custom Vision service should be aware of Microsoft's policies on customer data. 请参阅 Microsoft 信任中心上的“认知服务”页面来了解详细信息。See the Cognitive Services page on the Microsoft Trust Center to learn more.

后续步骤Next steps

按照生成分类器指南操作,开始在 Web 门户上使用自定义视觉,或者完成快速入门来实现代码中的基本方案。Follow the Build a classifier guide to get started using Custom Vision on the web portal, or complete a quickstart to implement the basic scenarios in code.