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.

• ✗

  Image captioning

  Why it's wrong here

  Image captioning is a computer vision technique that generates a human-readable sentence describing the visual content of an image, such as "a man holding a cup" or "a room with wooden furniture." It relies on object, scene, and action recognition to produce high-level semantic summaries. Crucially, it does not perform pixel-level character extraction, so it cannot isolate or transcribe the actual letters and digits that make up handwritten text. Even if a label is described, the handwritten characters themselves are never decoded into machine-readable text.

• ✓

  Optical Character Recognition (OCR)

  Why this is correct

  Optical Character Recognition (OCR), specifically the Azure AI Vision Read API, is purpose-built to extract text from images, including hard-to-read handwritten labels. It uses deep learning models trained on both printed and cursive scripts to identify individual characters, words, and lines, and it returns the recognized text along with bounding-box coordinates and confidence scores. This directly matches the museum's requirement to transcribe handwritten labels into digital, searchable text. OCR handles variations in handwriting style, angle, and background noise that would be impossible for other vision techniques.

• ✗

  Facial recognition

  Why it's wrong here

  Facial recognition is a specialized biometric technique that detects human faces and identifies or verifies individuals by comparing facial features like eye spacing, nose shape, and jawline. It is entirely concerned with face geometry and identity, and its models are trained on facial landmarks, not on alphanumeric characters or handwritten strokes. The museum's labels could contain a person's name, but facial recognition cannot read those letters; it can only recognize the face of the person appearing in the image. Therefore, this option provides no mechanism for extracting text from the labels.

• ✗

  Object detection

  Why it's wrong here

  Object detection is a computer vision task that locates and classifies objects within an image, returning bounding boxes and category labels such as "artifact," "display case," or "label." It can tell you that a label exists somewhere in the scene and where it is, but it does not read the characters written on that label. The algorithm compares image regions against learned object shapes and textures, not against letterforms or grapheme patterns. Consequently, object detection would fail to produce any transcribed text content for the museum's handwritten labels.

Go deeper

Related to this question