Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Important
Microsoft is announcing the planned retirement of the Azure Custom Vision service. Microsoft will provide full support for all existing Azure Custom Vision customers until 9/25/2028. During this support window, customers are encouraged to begin planning and executing their transition to alternative solutions. Depending on your use case, we recommend the following paths for transition:
- For creating custom models for both image classification and object detection, Azure Machine Learning AutoML offers the ability to train both custom model types using classic machine learning techniques
- Learn more about Azure Machine Learning AutoML and explore how it can offer support for custom model training.
Microsoft is also investing in Generative AI-based solutions that increase accuracy in custom scenarios using prompt engineering and other techniques.
- To use generative models, you can use one of models available in the Azure AI Foundry model catalog and create your own solution for customized vision.
- For a managed generative solution for image classification, Azure AI Content Understanding (currently in public preview) offers the ability to create custom classification workflows. It also supports processing unstructured data of any type (image, documents, audio, video) and extract structured insights based on pre-defined or user-defined formats.
- Learn more about Azure AI Foundry Models and Azure AI Content Understanding (public preview) and explore how they can offer alternative paths for your custom needs.
For more detailed guidance on migration, see the Azure Custom Vision Migration Guide.
Azure AI Custom Vision is an image recognition service that lets you build, deploy, and improve your own image identifier models. An image identifier applies labels to images according to their visual characteristics. Each label represents a classification or object. Custom Vision allows you to specify your own labels and train custom models to detect them.
You can use Custom Vision through a client library SDK, REST API, or through the Custom Vision web portal. Follow a quickstart to get started.
This documentation contains the following types of articles:
- The quickstarts are step-by-step instructions that let you make calls to the service and get results in a short period of time.
- The how-to guides contain instructions for using the service in more specific or customized ways.
For a more structured approach, follow a Training module for Custom Vision:
How it works
The Custom Vision service uses a machine learning algorithm to analyze images for custom features. You submit sets of images that do and don't have the visual characteristics you're looking for. Then you label the images with your own labels (tags) at the time of submission. The algorithm trains to this data and calculates its own accuracy by testing itself on the same images. Once you've trained your model, you can test, retrain, and eventually use it in your image recognition app to classify images or detect objects. You can also export the model for offline use.
Classification and object detection
Custom Vision functionality can be divided into two features. Image classification applies one or more labels to an entire image. Object detection is similar, but it returns the coordinates in the image where the applied label(s) are found.
Use case optimization
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. It's generally a good start to use 50 images per label. However, the service isn't optimal for detecting subtle differences in images (for example, detecting minor cracks or dents in quality assurance scenarios).
Additionally, you can choose from several variations of the Custom Vision algorithm that are optimized for images with certain subject material—for example, landmarks or retail items. For more information, see Select a domain.
How to use Custom Vision
The Custom Vision Service is available as a set of native SDKs and through a web-based interface on the Custom Vision portal. You can create, test, and train a model through either interface or use both together.
Supported browsers
The Custom Vision portal can be used by the following web browsers:
- Microsoft Edge (latest version)
- Google Chrome (latest version)
Backup and disaster recovery
As a part of Azure, Custom Vision Service has components that are maintained across multiple regions. Service zones and regions are used by all of our services to provide continued service to our customers. For more information on zones and regions, see Azure regions. If you need additional information or have any issues, contact support.
Input requirements
See Limits and quotas for image input limitations.
Data privacy and security
As with all of the Azure AI services, developers using the Custom Vision service should be aware of Microsoft's policies on customer data. See the Azure AI services page on the Microsoft Trust Center to learn more.
Data residency
Custom Vision doesn't replicate data outside of the specified region, except for one region, NorthCentralUS, where there is no local Azure Support.
Next steps
Follow the Build a classifier quickstart to get started using Custom Vision in the web portal.
- Or, complete an SDK quickstart to implement the basic scenarios with code.