Module 01 Knowledge Check
5 questions. Pass mark 4 out of 5. Answer every question before checking the answer key below, then retry after reading the feedback.
1. A colour image array has the shape…
- A. (width, height, channels)
- B. (height, width, channels)
- C. (channels, width, height)
- D. (width, channels, height)
2. Loading a BGR image and feeding it to an RGB-trained model…
- A. raises an error
- B. runs, and silently degrades every prediction
- C. auto-corrects
- D. only affects display
3. Publicly visible images are…
- A. free to train on
- B. not necessarily licensed for training
- C. usable if you credit the source
- D. public domain
4. EXIF metadata should be…
- A. kept for provenance
- B. stripped at ingestion
- C. kept only for training images
- D. ignored
5. The most durable field in a dataset card is…
- A. the image count
- B. what the dataset may NOT be used for
- C. the collection date
- D. the file format
---
Answer Key and Explanations
Check these only after attempting every question.
1. B - (height, width, channels)
numpy indexes rows before columns, which is transposed from the (x, y) most drawing APIs take.
2. B - runs, and silently degrades every prediction
The most common silent bug in vision code: no symptom other than being worse.
3. B - not necessarily licensed for training
Permission travels with every model trained on the data, and a weights file carries no provenance.
4. B - stripped at ingestion
It carries location and timestamp on every file - a movement record of whoever took the photographs.
5. B - what the dataset may NOT be used for
Models get reused, and the reuse is always a reasonable-sounding extension proposed by someone who was not there.
Practical Check
Apply this module to your own work: complete the module activity for *Images as Arrays and Visual Data Ethics*, then write one sentence naming what your result shows and one naming what it does not.
Strong Answer Pattern
A strong answer names the task, the evidence used, the check performed, and the remaining limitation. It avoids "proved", "guaranteed", or "always" unless the evidence genuinely supports it.