Microsoft Azure AI Fundamentals (AI-900)Describe features of computer vision workloads on AzureHard
A security company is developing a system to monitor restricted areas. They need to analyze video streams to detect specific types of unauthorized objects, such as drones or unfamiliar vehicles, as they enter the area. The system should also track their movement and alert security personnel. Which combination of computer vision capabilities would be most crucial for this system?
- AFacial Analysis and Face Recognition
- BObject Detection and Object Tracking
- COptical Character Recognition and Semantic Segmentation
- DImage Classification and Image Tagging
Show answer & explanationAnswer & explanation
Correct answer: B. Object Detection and Object Tracking
Object Detection identifies and locates the 'unauthorized objects' (drones, vehicles) in each frame. Object Tracking then follows the movement of these detected objects across consecutive frames, enabling alerts for their entry and movement, making this combination most crucial.
Why the other options are wrong
- A. Facial Analysis and Face Recognition are for human face attributes and identity, not general unauthorized objects like drones or vehicles.
- C. Optical Character Recognition extracts text, and Semantic Segmentation classifies pixels into categories; neither tracks moving objects.
- D. Image Classification categorizes the whole image, and Image Tagging adds keywords; neither provides location or movement tracking of specific objects.
Object Detection & Tracking
Object Detection identifies and locates objects in a single frame, while Object Tracking follows the movement of those identified objects across a sequence of frames in a video.
- Detection provides 'what' and 'where' at a moment in time.
- Tracking provides 'how it moves' over time.
- Often combined for surveillance, autonomous driving, and activity analysis.
Memory trick: Dynamic objects are first detected, then their path is tracked.