Microsoft Azure AI Fundamentals (AI-900) Practice Exam

When an image is sent to a Computer Vision API and receives an annotated image in return, what type of computer vision is typically used?

1. Facial recognition

2. Object detection

3. Optical character recognition (OCR)

4. Classification

The correct answer is: Object detection

The use of object detection in the context of receiving an annotated image from a Computer Vision API is significant because object detection involves identifying and locating multiple objects within an image, while also typically providing bounding boxes around those objects and associated labels. When an annotated image is returned, it often contains visual cues like labels, bounding boxes, or other metadata that describe the objects present in the original image. This process allows users to understand not just what objects are present in the image, but also where they are located, which is a core requirement of object detection tasks. In contrast, while facial recognition focuses specifically on identifying and verifying human faces, and optical character recognition (OCR) is aimed at reading and translating text from images, neither of these techniques captures the same breadth of information regarding various objects within an image as object detection does. Classification, on the other hand, typically assigns a single label to an entire image, not providing the detailed spatial information found in annotated images produced via object detection. Thus, object detection is the most appropriate and relevant type of computer vision utilized in this scenario.