AI in Medical Imaging Quiz

Answers and explanations appear as you take the quiz.

1. Which type of neural network has been most foundational for analyzing medical images?
   • Convolutional neural networks (CNNs)
   • Recurrent networks for audio
   • Decision trees
   • Spreadsheet macros
2. What does the Viz.ai system do?
   • Schedules hospital cafeteria menus
   • Analyzes CT scans to flag suspected large-vessel strokes and alert the team
   • Translates radiology reports into music
   • Replaces the CT scanner hardware
3. What is the U-Net architecture commonly used for in imaging?
   • Generating song lyrics
   • Compressing video games
   • Per-pixel segmentation, like outlining a tumor or organ
   • Encrypting patient passwords
4. What regulatory pathway did FDA use for the named Viz.ai imaging device?
   • A blanket approval covering all imaging-AI tools
   • A drug marketing authorization
   • An exemption from all clinical evidence requirements
   • A De Novo classification decision for a specific device and intended use
5. A key reason an imaging AI might fail on a new hospital's data is:
   • Differences in scanner type, protocol, or patient population from training data
   • The scans are too colorful
   • The model dislikes the building
   • Images are always identical everywhere
6. What do U-Net skip connections carry from the encoder to the decoder?
   • A patient’s medical record number
   • Higher-resolution feature maps that help recover spatial detail
   • A radiology report’s billing code
   • Only a global class label
7. Why should medical-imaging datasets generally be split at the patient level?
   • To make every scanner appear in only one dataset
   • To guarantee equal disease prevalence
   • To reduce leakage from multiple correlated images of the same patient across train and test sets
   • To remove the need for external validation
8. What does probability calibration assess in an imaging model?
   • Whether the scanner has been recently serviced
   • Whether every image has the same resolution
   • Whether the model segments the correct number of organs
   • Whether predictions assigned a probability match observed frequencies over evaluated cases