Skip to content

Commit 79df765

Browse files
PatrickFarleyPatrickFarley
committed
Merge branch 'patch-19' of https://github.com/kbrowne8/azure-ai-docs-pr into cusvis-retirement
2 parents cda1abf + 0fc77bc commit 79df765

File tree

3 files changed

+147
-0
lines changed

3 files changed

+147
-0
lines changed

‎articles/ai-services/computer-vision/migration-optionss.md‎

Lines changed: 70 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,70 @@
1+
# Migration Guidance for Transitioning from Computer Vision - Image Analysis
2+
3+
The Azure Computer Vision - Image Analysis API will be retired on September 25, 2028, after which calls made to the service will fail. Microsoft will provide full support for all existing Image Analysis customers until 9/25/2028, but to ensure business continuity and minimize disruption, customers are encouraged to begin planning their migration to alternative solutions that best meet their scenario requirements. This document provides comprehensive guidance for evaluating, selecting, and transitioning to new services.
4+
5+
## Migration Preparation Checklist
6+
1. Assess current usage and dependencies on Image Analysis API.
7+
2. Identify business scenarios and technical requirements for your image analysis scenarios.
8+
3. Evaluate alternative solutions based on capabilities, integration, cost, and support.
9+
4. Plan model migration steps.
10+
5. Test new solution(s) in a staging environment.
11+
6. Update production workflows and retrain stakeholders.
12+
13+
## Alternative Options Based on Scenario Needs
14+
There are several alternative platforms and services that can be considered depending on your specific use case, technical requirements, and integration needs. The following options are recommended for each set of features under Image Analysis.
15+
16+
### For OCR and Read capabilities try Document Intelligence
17+
18+
The Document Intelligence service provides support for OCR text in images.
19+
* **Features**: Azure AI Document Intelligence is a cloud-based Azure AI service that you can use to build intelligent document processing solutions.
20+
* **Learn more** about Document Intelligence:
21+
* [What is Azure AI Document Intelligence?]()
22+
* [Document Intelligence Read model]()
23+
24+
### For Face scenarios: try the Face API
25+
26+
The Face service offers Face detection capabilities, as well as a more comprehensive portfolio of face-related features.
27+
* **Features**: Full support for all Face scenarios under the Image Analysis API.
28+
* **Learn more** about the Face API:
29+
* [What is the Azure AI Face Service?]()
30+
* [Face detection, attributes, and input data]()
31+
32+
### For image embeddings scenarios:
33+
34+
#### Cohere Embed v3 in the Azure AI Foundry
35+
* **Best for**: Customers who need image + text embeddings supported on Azure.
36+
* **Features**: A multilingual multimodal embedding model supported in the Azure AI Foundry portal. It is capable of transforming different modalities such as images, texts, and interleaved images and texts into a single vector representation.
37+
* **Learn more** about Cohere Embed v4:
38+
* [Embed-v-4-0]()
39+
40+
#### SigLIP (Sigmoid Loss for Language Image Pre training)
41+
* **Best for**: Customers who need strong zero shot classification and image text retrieval abilities.
42+
* **Features**: A CLIP‐style vision‐language model from Google that replaces the standard contrastive (softmax) loss with a pairwise sigmoid loss. It trains on large scale image text pairs.
43+
* **Learn more** about SigLIP:
44+
* [Sigmoid Loss for Language Image Pre-Training]()
45+
* [SigLP on Hugging Face]()
46+
47+
### For other Computer Vision scenarios:
48+
49+
There are multiple additional alternative services that can support the remaining scenarios supported in the Image Analysis API.
50+
51+
#### GPT Model Series in the Azure AI Foundry
52+
53+
* **Best for**: Customers who are flexible in their approach to creating a solution for customized vision capabilities.
54+
* **Features**: Flexibility to build custom solutions based on different Generative AI models.
55+
* **Learn more** about Generative AI models in the Azure AI Foundry:
56+
* [Explore Azure AI Foundry Models]()
57+
* [Azure OpenAI in Azure AI Foundry models]()
58+
59+
#### Azure AI Content Understanding (preview)
60+
* **Best for**: Customers wanting a managed generative solution for image analysis scenarios.
61+
* **Features**: Content Understanding supports processing unstructured image data, as well as documents, audio, and video. It enables you to extract structured insights based on pre-defined or user-defined formats.
62+
* **Learn more** about Content Understanding:
63+
* [What is Azure AI Content Understanding?]()
64+
* [Azure AI Content Understanding image solutions (preview)]()
65+
* [Content Understanding classifier]()
66+
67+
## Next steps & required actions
68+
* Make a plan to transition away from Azure Computer Vision – Image Analysis by September 25, 2026
69+
* Azure Computer Vision – Image Analysis will be retired on 25 September 2028, please transition to alternative options by that date
70+

‎articles/ai-services/custom-vision-service/includes/custom-vision-retirement.md‎

Lines changed: 25 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,25 @@
1+
---
2+
title: "Custom Vision Services Retirement"
3+
titleSuffix: Azure AI services
4+
author: PatrickFarley
5+
manager: nitinme
6+
ms.service: azure-ai-custom-vision
7+
ms.custom: build-2023
8+
ms.topic: include
9+
ms.date: 09/04/2025
10+
ms.author: pafarley
11+
---
12+
13+
14+
> [!IMPORTANT]
15+
> 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.
16+
> Depending on your use case, we recommend the following paths for transition:
17+
> - 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
18+
> - [Learn more about Azure Machine Learning AutoML]() and explore how it can offer support for custom model training.
19+
>
20+
> Microsoft is also investing in Generative AI-based solutions that increase accuracy in custom scenarios using prompt engineering and other techniques.
21+
> - 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.
22+
> - 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.
23+
> - [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.
24+
>
25+
> For more detailed guidance on migration, see the (Azure Custom Vision Migration Guide)[https://aka.ms/custom-vision-migration].

‎articles/ai-services/custom-vision-service/migration-options.md‎

Lines changed: 52 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,52 @@
1+
# Migration Guidance for Transitioning from Custom Vision Service
2+
3+
The Custom Vision Service will be retired on September 25, 2028, after which calls made to the service will fail. Microsoft will provide full support for all existing Azure Custom Vision customers until 9/25/2028, but to ensure business continuity and minimize disruption, customers are encouraged to begin planning their migration to alternative solutions that best meet their scenario requirements. This document provides comprehensive guidance for evaluating, selecting, and transitioning to new services.
4+
5+
## Migration Preparation Checklist
6+
7+
1. Assess current usage and dependencies on Custom Vision Service.
8+
2. Identify business scenarios and technical requirements for image classification and object detection.
9+
3. Evaluate alternative solutions based on capabilities, integration, cost, and support.
10+
4. Plan data export and model migration steps.
11+
5. Test new solution(s) in a staging environment.
12+
6. Update production workflows and retrain stakeholders.
13+
14+
## Alternative Options Based on Scenario Needs
15+
There are several alternative platforms and services that can be considered depending on your specific use case, technical requirements, and integration needs. The following options are recommended for common scenarios:
16+
17+
### Traditional Machine Learning options
18+
To create both custom image classification and object detection models using traditional machine learning techniques, take a look at the following option:
19+
20+
#### Azure Machine Learning AutoML
21+
* **Best for**: Customers seeking to apply classic machine learning techniques
22+
* **Features**: Offers a code-first experience, as well as a no-code studio web experience similar to Custom Vision. It offers the ability to easily train custom image classification and object detection models on your image data.
23+
* **Learn more** about Azure Machine Learning AutoML:
24+
* [What is automated machine learning?]()
25+
* [Set up no-code Automated ML training for tabular data with the studio UI]()
26+
* [Set up AutoML to train computer vision models]()
27+
28+
### Generative AI-based solutions
29+
Microsoft is also investing in Generative AI-based solutions that increase accuracy in custom scenarios using prompt engineering and other techniques.
30+
31+
#### Generative AI solutions in the Azure AI Foundry
32+
33+
* **Best for**: Customers who are flexible in their approach to creating a solution for customized vision capabilities.
34+
* **Features**: Flexibility to build custom solutions based on different Generative AI models.
35+
* **Learn more** about Generative AI models in the Azure AI Foundry:
36+
* [Explore Azure AI Foundry Models]()
37+
* [Azure OpenAI in Azure AI Foundry models]()
38+
39+
#### Azure AI Content Understanding (preview)
40+
* **Best for**: Customers wanting a managed generative solution for image classification
41+
* **Features**: Content Understanding offers the ability to create custom classification workflows. It also supports processing unstructured data of any type (image, documents, audio, video) and extracting structured insights based on pre-defined or user-defined formats.
42+
* **Learn more** about Content Understanding:
43+
* [What is Azure AI Content Understanding?]()
44+
* [Azure AI Content Understanding image solutions (preview)]()
45+
* [Content Understanding classifier]()
46+
47+
## Data Migration Guidance
48+
Before migrating, export your labeled datasets and model metadata from Custom Vision Service. Review the data formats required by your chosen alternative and convert as needed.
49+
50+
## Next steps & required actions
51+
* Make a plan to transition away from Azure Custom Vision by September 25, 2026
52+
* Azure Custom Vision will be retired on 25 September 2028, please transition to alternative options by that date.

0 commit comments

Comments
 (0)