Microsoft Azure AI Fundamentals (AI-900)Describe fundamental principles of machine learning on AzureMedium
A research team is developing an AI system to analyze medical images (e.g., X-rays, MRIs) to automatically detect and delineate tumors or other anomalies. The system needs to precisely identify the boundaries of these regions within the image. Which computer vision technique is most suitable for this task?
- AImage Segmentation
- BImage Classification
- CFeature Extraction
- DObject Detection
Show answer & explanationAnswer & explanation
Correct answer: A. Image Segmentation
Image segmentation is the most suitable technique as it involves partitioning an image into multiple segments (sets of pixels), specifically identifying and delineating the precise boundaries of objects like tumors.
Why the other options are wrong
- B. Image classification assigns a single label to the entire image (e.g., 'tumor present' or 'no tumor'), but doesn't locate or delineate it.
- C. Feature extraction is a preprocessing step to derive descriptive attributes from images, not a task for delineating objects itself.
- D. Object detection draws bounding boxes around objects, locating them but not precisely outlining their boundaries.
Image Segmentation
A computer vision technique that partitions a digital image into multiple segments (sets of pixels), often to locate objects and boundaries.
- Assigns a label to every pixel in an image.
- Provides precise object boundaries (pixel-level understanding).
- Used in medical imaging, autonomous driving, satellite imagery analysis.
Memory trick: Vision Tasks: See, Locate, Outline, Understand