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.

• ✗

  Estimating the correct posture for employees based on ergonomics guidelines

  Why it's wrong here

  Estimating the correct posture for employees according to ergonomics guidelines is an evaluation task that compares observed body alignment to external health and safety standards, not a low-level vision task. Human pose estimation does not make normative judgments; it simply localizes anatomical keypoints such as elbows, knees, and shoulders in image coordinates. Any ergonomics-based assessment would be a downstream decision-making layer built on top of the pose data, not the core vision capability being described.

• ✓

  Detecting body keypoint positions (joints) in images to infer posture and movement

  Why this is correct

  Pose estimation is a computer vision technique that uses neural network regression or heatmap-based decoding to predict the 2D or 3D coordinates of body joints (keypoints) from an image. These keypoints, often following a skeletal model like COCO's 17-joint layout, allow algorithms to reconstruct the subject's posture and track movement across frames. This output underpins applications such as fitness coaching, motion capture, and gesture recognition.

• ✗

  Determining the camera angle and position used to capture a photograph

  Why it's wrong here

  Determining the camera angle and position that captured a photograph refers to camera pose estimation, which solves for the camera's extrinsic parameters—its rotation and translation relative to the scene—using geometric feature correspondences. Human pose estimation is frequently confused with this term because both share the word 'pose,' but it operates on the human body, not the camera. The former identifies where joints appear in the image; the latter estimates where the camera was located in the real world.

• ✗

  Classifying whether a person is sitting or standing in an image

  Why it's wrong here

  Classifying a person as sitting or standing is a scene-level action recognition problem, typically solved with an image classifier that assigns a single semantic label to the entire image. Unlike pose estimation, it does not output any spatial information about individual joints or limb orientations. The label could be inferred from contextual cues like furniture or object relationships rather than an articulated body model, making it a coarser, higher-level abstraction than joint-level pose estimation.

Go deeper

Related to this question