Merge branch 'MicrosoftDocs:main' into main

Author: juliamuiruri4

articles/ai-foundry/model-inference/quotas-limits.md

@@ -17,7 +17,7 @@ This article contains a quick reference and a detailed description of the quotas
 
 ## Quotas and limits reference
 
-The following sections provide you with a quick guide to the default quotas and limits that apply to Azure AI model's inference service in Azure AI services:
+Azure uses quotas and limits to prevent budget overruns due to fraud, and to honor Azure capacity constraints. Consider these limits as you scale for production workloads. The following sections provide you with a quick guide to the default quotas and limits that apply to Azure AI model's inference service in Azure AI services:
 
 ### Resource limits
 
@@ -28,12 +28,18 @@ The following sections provide you with a quick guide to the default quotas and
 
 ### Rate limits
 
-| Limit name | Limit value |
-| ---------- | ----------- |
-| Tokens per minute (Azure OpenAI models)   | Varies per model and SKU. See [limits for Azure OpenAI](../../ai-services/openai/quotas-limits.md). |
-| Tokens per minute (rest of models)        | 200.000 |
-| Requests per minute (Azure OpenAI models) | Varies per model and SKU. See [limits for Azure OpenAI](../../ai-services/openai/quotas-limits.md). |
-| Requests per minute (rest of models)      | 1.000   |
+| Limit name           | Applies to          | Limit value |
+| -------------------- | ------------------- | ----------- |
+| Tokens per minute    | Azure OpenAI models | Varies per model and SKU. See [limits for Azure OpenAI](../../ai-services/openai/quotas-limits.md). |
+| Requests per minute  | Azure OpenAI models | Varies per model and SKU. See [limits for Azure OpenAI](../../ai-services/openai/quotas-limits.md). |
+| Tokens per minute    | DeepSeek models     | 5.000.000 |
+| Requests per minute  | DeepSeek models     | 5.000     |
+| Concurrent requests  | DeepSeek models     | 300        |
+| Tokens per minute    | Rest of models      | 200.000   |
+| Requests per minute  | Rest of models      | 1.000     |
+| Concurrent requests  | Rest of models      | 300       |
+
+You can [request increases to the default limits](#request-increases-to-the-default-limits). Due to high demand, limit increase requests can be submitted and evaluated per request.
 
 ### Other limits
 
@@ -49,6 +55,28 @@ Global Standard deployments use Azure's global infrastructure, dynamically routi
 
 The Usage Limit determines the level of usage above which customers might see larger variability in response latency. A customer's usage is defined per model and is the total tokens consumed across all deployments in all subscriptions in all regions for a given tenant.
 
+## Request increases to the default limits
+
+Limit increase requests can be submitted and evaluated per request. [Open an online customer support request](https://portal.azure.com/#blade/Microsoft_Azure_Support/HelpAndSupportBlade/newsupportrequest/). When requesting for endpoint limit increase, provide the following information:
+
+1. When opening the support request, select **Service and subscription limits (quotas)** as the **Issue type**.
+
+1. Select the subscription of your choice.
+
+1. Select **Cognitive Services** as **Quota type**.
+
+1. Select **Next**.
+
+1. On the **Additional details** tab, you need to provide detailed reasons for the limit increase in order for your request to be processed. Be sure to add the following information into the reason for limit increase:
+
+   * Model name, model version (if applicable), and deployment type (SKU).
+   * Description of your scenario and workload.
+   * Rationale for the requested increase.
+   * Provide the target throughput: Tokens per minute, requests per minute, etc.
+   * Provide planned time plan (by when you need increased limits).
+
+1. Finally, select **Save and continue** to continue.
+
 ## General best practices to remain within rate limits
 
 To minimize issues related to rate limits, it's a good idea to use the following techniques:
@@ -58,10 +86,6 @@ To minimize issues related to rate limits, it's a good idea to use the following
 - Test different load increase patterns.
 - Increase the quota assigned to your deployment. Move quota from another deployment, if necessary.
 
-### Request increases to the default quotas and limits
-
-Quota increase requests can be submitted and evaluated per request. [Submit a service request](../../ai-services/cognitive-services-support-options.md?context=/azure/ai-services/openai/context/context).
-
 ## Next steps
 
-* Learn more about the [models available in the Azure AI model's inference service](./concepts/models.md)
+* Learn more about the [models available in the Azure AI model's inference service](./concepts/models.md)

articles/ai-services/computer-vision/concept-face-liveness-detection.md

@@ -0,0 +1,89 @@
+---
+title: "Face liveness detection - Face"
+titleSuffix: Azure AI services
+description: This article explains the concept of Face liveness detection, its input and output schema, and related concepts. 
+author: PatrickFarley
+manager: nitinme
+
+ms.service: azure-ai-vision
+ms.subservice: azure-ai-face
+ms.custom:
+ms.topic: conceptual
+ms.date: 02/19/2025
+ms.author: pafarley
+feedback_help_link_url: https://learn.microsoft.com/answers/tags/156/azure-face
+---
+
+# Face liveness detection
+
+This article explains the concept of Face liveness detection, its input and output schema, and related concepts.
+
+## Introduction
+
+Face Liveness detection is used to determine if a face in an input video stream is real (live) or fake (spoofed). It's an important building block in a biometric authentication system to prevent imposters from gaining access to the system using a photograph, video, mask, or other means to impersonate another person.
+
+The goal of liveness detection is to ensure that the system is interacting with a physically present, live person at the time of authentication. These systems are increasingly important with the rise of digital finance, remote access control, and online identity verification processes.
+
+The Azure AI Face liveness detection solution successfully defends against various spoof types ranging from paper printouts, 2D/3D masks, and spoof presentations on phones and laptops. Liveness detection is an active area of research, with continuous improvements being made to counteract increasingly sophisticated spoofing attacks. Continuous improvements are rolled out to the client and the service components over time as the overall solution gets more robust to new types of attacks.
+
+The Azure Face liveness detection API is [conformant to ISO/IEC 30107-3 PAD (Presentation Attack Detection) standards](https://www.ibeta.com/wp-content/uploads/2023/11/230622-Microsoft-PAD-Level-2-Confirmation-Letter.pdf) as validated by iBeta level 1 and level 2 conformance testing.
+
+## How it works
+
+The liveness solution integration involves two distinct components: a frontend mobile/web application and an app server/orchestrator.
+
+:::image type="content" source="./media/liveness/liveness-diagram.jpg" alt-text="Diagram of the liveness workflow in Azure AI Face." lightbox="./media/liveness/liveness-diagram.jpg":::
+
+- **Frontend application**: The frontend application receives authorization from the app server to initiate liveness detection. Its primary objective is to activate the camera and guide end-users accurately through the liveness detection process.
+- **App server**: The app server serves as a backend server to create liveness detection sessions and obtain an authorization token from the Face service for a particular session. This token authorizes the frontend application to perform liveness detection. The app server's objectives are to manage the sessions, to grant authorization for frontend application, and to view the results of the liveness detection process.
+
+
+## Liveness detection modes
+
+Azure Face liveness detection API includes options for both Passive and Passive-Active detection modes.
+
+The **Passive mode** utilizes a passive liveness technique that requires no additional actions from the user. It requires a non-bright lighting environment to succeed and will fail in bright lighting environments with an "Environment not supported" error. It also requires high screen brightness for optimal performance which is configured automatically in the Mobile (iOS and Android) solutions. This mode can be chosen if you prefer minimal end-user interaction and expect end-users to primarily be in non-bright environments. A Passive mode check takes around 12 seconds on an average to complete.
+
+The **Passive-Active mode** will behave the same as the Passive mode in non-bright lighting environments and only trigger the Active mode in bright lighting environments. This mode is preferable on Web browser solutions due to the lack of automatic screen brightness control available on browsers which hinders the Passive mode's operational envelope. This mode can be chosen if you want the liveness-check to work in any lighting environment. If the Active check is triggered due to a bright lighting environment, then the total completion time may take up to 20 seconds on average.
+
+You can set the detection mode during the session creation step (see [Perform liveness detection](./tutorials/liveness.md#perform-liveness-detection)).
+
+## Optional face verification
+
+You can combine face verification with liveness detection to verify that the face in question belongs to the particular person designated. The following table describes details of the liveness detection features:
+
+| Feature | Description |
+| -- |--|
+| Liveness detection | Determine an input is real or fake, and only the app server has the authority to start the liveness check and query the result. |
+| Liveness detection with face verification | Determine an input is real or fake and verify the identity of the person based on a reference image you provided. Either the app server or the frontend application can provide a reference image. Only the app server has the authority to initial the liveness check and query the result. |
+
+
+
+
+## Output format
+
+The liveness detection API returns a JSON object with the following information:
+- A Real or a Spoof Face Liveness Decision. We handle the underlying accuracy and thresholding, so you don’t have to worry about interpreting “confidence scores” or making inferences yourself. This makes integration easier and more seamless for developers.
+- Optionally a Face Verification result can be obtained if the liveness check is performed with verification (see [Perform liveness detection with face verification](./tutorials/liveness.md#perform-liveness-detection-with-face-verification)).
+- A quality filtered "session-image" that can be used to store for auditing purposes or for human review or to perform further analysis using the Face service APIs.
+
+
+## Data privacy
+
+We do not store any images or videos from the Face Liveness Check. No image/video data is stored in the liveness service after the liveness session has been concluded. Moreover, the image/video uploaded during the liveness check is only used to perform the liveness classification to determine if the user is real or a spoof (and optionally to perform a match against a reference image in the liveness-with-verify-scenario), and it cannot be viewed by any human and will not be used for any AI model improvements.
+
+## Security
+
+We include additional runtime application self-protections (RASP), provided by [GuardSquare]( https://www.guardsquare.com/blog/why-guardsquare), in our Mobile SDKs (iOS and Android). 
+
+## Support options
+
+In addition to using the main [Azure AI services support options](../cognitive-services-support-options.md), you can also post your questions in the [issues](https://github.com/Azure-Samples/azure-ai-vision-sdk/issues) section of the SDK repo. 
+
+
+## Next step
+
+Now that you're familiar with liveness detection concepts, implement liveness detection in your app.
+
+> [!div class="nextstepaction"]
+> [Face liveness detection](./tutorials/liveness.md)

articles/ai-services/computer-vision/concept-liveness-abuse-monitoring.md

@@ -9,7 +9,7 @@ manager: nitinme
 ms.service: azure-ai-vision
 ms.subservice: azure-ai-face
 ms.topic: conceptual
-ms.date: 01/29/2025
+ms.date: 02/22/2025
 ms.author: pafarley
 ms.custom:
   - ignite-2023
@@ -20,7 +20,7 @@ feedback_help_link_url: https://learn.microsoft.com/answers/tags/156/azure-face
 
 Azure AI Face liveness detection lets you detect and mitigate instances of recurring content and/or behaviors that indicate a violation of the [Code of Conduct](/legal/cognitive-services/face/code-of-conduct?context=/azure/ai-services/computer-vision/context/context) or other applicable product terms. This guide shows you how to work with these features to ensure your application is compliant with Azure policy.
 
-Details on how data is handled can be found on the [Data, Privacy, and Security](/legal/cognitive-services/openai/data-privacy?context=/azure/ai-services/openai/context/context) page.
+Details on how data is handled can be found on the [Data, Privacy, and Security](/legal/cognitive-services/face/data-privacy-security?context=/azure/ai-services/computer-vision/context/context) page.
 
 [!INCLUDE [liveness-sdk-gate](./includes/liveness-sdk-gate.md)]
 

articles/ai-services/computer-vision/how-to/mitigate-latency.md

@@ -72,7 +72,6 @@ The quality of the input images affects both the accuracy and the latency of the
 
 To achieve the optimal balance between accuracy and speed, follow these tips to optimize your input data. 
 - For face detection and recognition operations, see [input data for face detection](../concept-face-detection.md#input-requirements) and [input data for face recognition](../concept-face-recognition.md#input-requirements).
-- For liveness detection, see the [tutorial](../Tutorials/liveness.md#select-a-reference-image). 
 
 
 ## Call APIs in parallel when possible

articles/ai-services/computer-vision/includes/identity-input-verification.md

@@ -0,0 +1,23 @@
+---
+title: Face identity verification input recommendations
+titleSuffix: Azure AI services
+#services: cognitive-services
+author: PatrickFarley
+manager: nitinme
+ms.service: azure-ai-vision
+ms.custom:
+ms.topic: include
+ms.date: 02/23/2025
+ms.author: pafarley
+---
+
+- Photo is clear and sharp—not blurry, pixelated, distorted, or damaged.
+- Photo is not altered to remove face blemishes or alter face appearance.
+- Photo must be in an RGB color supported format (JPEG, PNG, WEBP, BMP). Recommended Face size is 200 px x 200 px. Face sizes larger than 200x200 won't result in better AI quality. Image files must not be larger than 6 MB in size.
+- User isn't wearing glasses, masks, hats, headphones, head coverings, or face coverings. Face should be free of any obstructions.
+- Facial jewelry is allowed provided it doesn't hide the face.
+- Only one face should be visible in the photo.
+- Face should be in a neutral front-facing pose with both eyes open, mouth closed, with no extreme facial expressions or head tilt.
+- Face should be free of any shadows or red-eye. Retake photo if either of these features appear.
+- Background should be uniform and plain, free of any shadows.
+- Face should be centered within the image and fill at least 50% of the image.

articles/ai-services/computer-vision/overview-identity.md

@@ -157,6 +157,8 @@ Input requirements for face detection:
 Input requirements for face recognition:
 [!INCLUDE [identity-input-composition](includes/identity-input-composition.md)]
 
+Input requirements for face verification:
+[!INCLUDE [identity-input-verification](includes/identity-input-verification.md)]
 
 ## Data privacy and security