AI-103: Computer Vision
Computer Vision is one of the skill areas tested on the Developing AI Apps & Agents on Azure (AI-103) exam. Below are free AI-103 practice questions with worked answers and a concept diagram — each with a plain-language explanation. Practice the first 10 questions of AI-103 free, no signup.
The concept, in one picture
2 free AI-103 Computer Vision questions
In a Microsoft Foundry project you are building an image-editing workflow that must satisfy these needs: remove background objects using a mask-based inpainting edit, keep the original lighting and style of the images intact, and rely on the built-in image-editing controls rather than a custom model. Edits must affect only the region inside the mask. Which configuration should you choose?
- ✓ Turn on mask_inpainting and pass both the source image and a mask that marks the region to change.
- Switch to image_to_image mode with a high strength value so the whole image is regenerated from the prompt.
- Use text_to_image mode with a prompt that describes removing the background objects.
- Choose image_variation mode and supply the original image as the reference.
The goal is a localized edit of an existing image that preserves the surrounding lighting, composition, and style, not the creation of a brand-new image. Azure OpenAI image editing in Microsoft Foundry edits an existing image from an input image plus a prompt, and for masked edits the mask specifies the exact area the model may alter (its dimensions must match the input image). text_to_image would create a fresh image with no guarantee of preserving the original. image_variation produces related variants instead of a targeted removal. image_to_image at high strength can regenerate large regions and change unrelated details. Mask-based inpainting is the built-in control that confines changes to the masked region.
A Microsoft Foundry project produces product marketing images from text prompts. After you publish several images, legal spots a competitor's logo on a sign in the background of one image. You must delete only that logo while leaving the rest of the image unchanged. What should you do?
- Raise the prompt guidance strength value.
- Edit the original prompt so it forbids brand names, then regenerate.
- ✓ Run a mask-based inpainting edit over the area of the image that holds the logo.
- Re-execute the prompt with a new random seed.
This is a targeted edit: remove only the competitor logo while keeping the rest of an already-produced image. Azure OpenAI image editing changes existing images from a text instruction instead of regenerating them, and the Image Edit API needs the input image in the request. With mask-based inpainting the mask marks the precise region to modify, so only the logo area is replaced and the surrounding background, composition, lighting, and product remain. Increasing prompt guidance strength only influences generation adherence and cannot strip a logo from a finished image. Rewriting the prompt and regenerating yields a different image, and a new seed changes the result unpredictably and may add new problems. Mask-based inpainting is the minimal-change fix for post-generation brand cleanup.
Practice AI-103 free
The first 10 questions of every exam are free. No signup, no email wall.
Start practicing →Get a free AI-103 study plan by email
A short plan to work through AI-103 by skill area, plus a note when we add new questions. Optional — the practice above stays free. No spam, unsubscribe anytime.