3.3. Reflection Checkpoint: Computer Vision Mastery

Before proceeding, ensure you can answer these questions:

1. What's the key difference between image classification and object detection?

   • Classification: One label for entire image
   • Object Detection: Multiple objects with bounding box coordinates

2. What does OCR return?

   • Text strings extracted from images (with position information)

3. Which Face service feature detects glasses and headwear?

   • Face Attributes

4. What specialized domain models does Azure Vision support?

   • Celebrities and Landmarks only

5. What three ML steps does Azure Vision eliminate?

   • Choosing, training, and evaluating models (pre-built)

6. When should you use Custom Vision instead of pre-built Image Analysis?

   • When detecting objects unique to your organization (your products, your defects, your logos) that wouldn't exist in general training data

7. A company needs to detect their proprietary product packaging on store shelves. Which service?

   • Custom Vision (proprietary/unique objects)