how to setup image tagging and object detection

justin vermeulen 0 Reputation points
2026-06-08T07:53:50.6866667+00:00

i have setup a computer vision resource in Azure but do not find image tagging or object detection in vision studio. Where can i find and set this up ?

Azure AI Custom Vision
Azure AI Custom Vision

An Azure artificial intelligence service and end-to-end platform for applying computer vision to specific domains.

0 comments

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. SRILAKSHMI C 19,725 Reputation points Microsoft External Staff Moderator
    2026-06-08T13:54:02.48+00:00

    Hello @justin vermeulen

    Thank you for reaching out to Microsoft Q&A.

    I understand that you have created an Azure AI Vision (Computer Vision/Image Analysis) resource and are expecting to see Image Tagging and Object Detection options directly within Vision Studio. This can sometimes be confusing because Azure provides both prebuilt vision capabilities and custom model training capabilities through different services.

    Understanding the Available Options

    Azure AI Vision (Image Analysis) provides prebuilt capabilities such as:

    • Image tagging
    • Object detection
    • Caption generation
    • OCR (text extraction)
    • Dense captions and visual feature extraction

    These capabilities do not require you to manually tag images or train a model. Instead, you typically analyze images using the Azure AI Vision/Image Analysis APIs or the available Vision Studio experiences.

    If Your Goal Is Prebuilt Object Detection

    You can use Azure AI Vision's object detection capabilities to identify common objects within images using the Image Analysis service. A few important considerations:

    • Supported image formats, file sizes, and dimension limits apply.
    • Detection accuracy may vary for very small, partially visible, or densely packed objects.
    • Some features or models may not be available in all regions.
    • If a particular model is not supported, the service may return a "NotSupportedModel" error.

    If Your Goal Is to Train Your Own Model

    If you are looking for a workflow where you:

    • Upload images
    • Draw bounding boxes around objects
    • Create custom tags/labels
    • Train a custom object detection model

    then you are likely looking for Custom Vision rather than the standard Azure AI Vision/Image Analysis service.

    Custom Vision provides the classic:

    Upload Images → Tag Objects → Train Model → Publish Model

    workflow for custom image classification and object detection scenarios.

    Why You May Not See Tagging or Training Options

    In many cases, users create a Computer Vision/Azure AI Vision resource and expect to find an interface for manually tagging images and training models. However, the tagging and training experience is associated with Custom Vision, while Azure AI Vision focuses primarily on prebuilt image analysis capabilities.

    Quick Check

    Could you confirm which scenario matches your requirement?

    A. Detect common objects using Microsoft's prebuilt models (no training required)

    → Use Azure AI Vision (Image Analysis) object detection.

    B. Train your own object detection model by tagging objects in your images

    → Use Custom Vision (upload, tag, train, and publish workflow).

    Please refer this

    Computer Vision - Object Detection: https://docs.microsoft.com/azure/ai-services/computer-vision/concept-object-detection

    What is Image Analysis?: https://docs.microsoft.com/azure/cognitive-services/computer-vision/overview-image-analysis

    Call the Image Analysis API: https://docs.microsoft.com/azure/cognitive-services/computer-vision/vision-api-how-to-topics/howtocallvisionapi

    Custom Vision overview (“what is Custom Vision”): https://learn.microsoft.com/azure/ai-services/custom-vision-service/overview

    Custom Vision quickstart (Build an object detector): https://learn.microsoft.com/azure/ai-services/custom-vision-service/get-started-build-detector

    Custom Vision example tagging steps (object detector training): https://learn.microsoft.com/windows/mixed-reality/develop/unity/tutorials/mr-azure-310?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#chapter-2---training-your-custom-vision-project

    I Hope this helps. Do let me know if you have any further queries.

    Thank you!

    Was this answer helpful?

  2. AI answer

    2026-06-08T07:54:14.6+00:00

    Image tagging and object detection are not part of the generic Computer Vision “Vision Studio” resource; they are provided by Azure AI Custom Vision.

    To set this up, use the Custom Vision service instead of (or in addition to) the Computer Vision resource:

    1. Create a Custom Vision resource in Azure
      • In the Azure portal, create an Azure AI Custom Vision resource (training and prediction) in a resource group.
    2. Use the Custom Vision website for image tagging / object detection
      • Go to the Custom Vision website and sign in with the same account used in the Azure portal.
      • On the My Projects page, select New Project.
      • Choose your Custom Vision training resource from the Resource dropdown.
      • For Project Types, pick:
        • Classification for image tagging (image classification), or
        • Object Detection for object detection.
      • Select an appropriate Domain (for object detection: General, Logo, Products on shelves, or a Compact domain if models need to run on devices).
      • Select Create project.
    3. Train and use the model
      • Upload and tag images in the project, then train the model and use the prediction endpoint from your app or via the SDKs/REST API.

    If a no‑code experience is preferred, the browser-based Custom Vision portal is the primary place to configure both image tagging (classification) and object detection; they will not appear as separate options inside Vision Studio for a plain Computer Vision resource.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.