AI-900 Practice Question: Describe features of computer vision workloads on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure AI Vision is for video; Custom Vision is for still images only

  Why it's wrong here

  This is incorrect because Azure AI Vision is not limited to video; its core Image Analysis API works on still images, and it also handles video by extracting frames for analysis. Custom Vision is designed for still images used in image classification and object detection, but it is not forced into a video-only role like the option implies. The real distinction remains that Azure AI Vision is pre-built and general-purpose, whereas Custom Vision is trained on your own domain-specific data.

• ✓

  Azure AI Vision offers pre-built general models; Custom Vision lets you train models for your specific categories

  Why this is correct

  Azure AI Vision provides immediate access to Microsoft's pre-trained, general-purpose image analysis models—such as OCR, object detection, image tagging, and landmark recognition—without any custom training required. Custom Vision, in contrast, lets you upload your own labeled images and train a model to recognize your specific categories, such as a particular defect or animal species. Therefore, the correct answer is that one offers out-of-the-box general models and the other offers user-customized classification or detection.

• ✗

  Custom Vision is more expensive because it uses more advanced AI algorithms

  Why it's wrong here

  Cost differences do not stem from Custom Vision using "more advanced algorithms"—instead, Custom Vision is priced differently because it includes a training phase (charged per training hour) plus prediction usage, while Azure AI Vision charges per API call for its already-trained general models. Both services rely on deep neural networks, and Custom Vision's value is not algorithmic superiority but the ability to fine-tune a base model to your domain through transfer learning. Choosing based solely on price misses the functional distinction: pre-built vs. custom-trained.

• ✗

  Azure AI Vision requires GPU compute; Custom Vision runs on CPU only

  Why it's wrong here

  This is false because compute requirements are not the defining factor when choosing between Azure AI Vision and Custom Vision. Both are managed cloud services, so you do not provision or select GPU resources; Custom Vision training runs on Microsoft-managed infrastructure that can use GPU accelerators on the backend. The real differentiator is whether you need a pre-built general model or a model custom-trained on your own labeled images.

Go deeper

Related to this question