1. Which task outputs bounding boxes with labels?
- Semantic segmentation
- Image classification
- Object detection
- OCR only
Show answer
Answer: Object detection
Detection localises objects with boxes.
Lesson: What Computer Vision Does
26 questions from the Computer Vision course, each with the correct answer and an explanation. Try to answer before you open the answer.
Answer: Object detection
Detection localises objects with boxes.
Lesson: What Computer Vision Does
Answer: (channels, height, width)
Convert with permute when moving from NumPy.
Lesson: Images as Arrays of Pixels
Answer: It separates colour (hue) from brightness, so masks are more robust to lighting
Hue stays similar when brightness changes.
Lesson: Colour Spaces
Answer: It reduces noise but also softens edges
Smoothing removes fine detail too.
Lesson: Filtering and Denoising
Answer: Fewer, stronger edges are kept
Higher thresholds keep only strong gradients.
Lesson: Edge Detection
Answer: Several touching coins merged into one connected component
Touching objects need separation methods.
Answer: Fits a geometric transform while rejecting outlier matches
Robust fitting ignores wrong matches.
Lesson: Keypoints and Feature Matching
Answer: They must be transformed with the same rotation
Labels must follow the pixels.
Lesson: Geometric Transforms
Answer: So evaluation scores are stable and comparable
Random changes belong in training.
Lesson: Data Augmentation
Answer: Near-duplicate frames inflate test scores
Keep related images on one side of the split.
Lesson: Datasets and Labelling
Answer: Residual (skip) connections
Skip connections keep gradients flowing.
Lesson: CNN Architectures in Brief
Answer: The model learned from inputs scaled and normalised that way
Mismatched inputs degrade predictions.
Answer: Only a simple classifier on top of frozen pretrained features
Cheap, fast and often strong.
Answer: 0.67
Overlap shrinks quickly with shifts.
Lesson: Bounding Boxes and IoU
Answer: Lower-scoring boxes that heavily overlap a higher-scoring box
It removes duplicates of the same object.
Lesson: Non-Maximum Suppression
Answer: Detectors return many low-confidence boxes that are usually wrong
Raw outputs include many weak guesses.
Lesson: Running a Pretrained Detector
Answer: Each ground-truth object can be matched only once
Duplicates are penalised.
Answer: Instance segmentation
Instances are individual objects.
Answer: Taking the argmax over class scores for each pixel
One class per pixel.
Answer: Dice is always greater than or equal to IoU
Do not compare one with the other.
Answer: Confidence can collapse even when the top label stays the same
Falling confidence is an early warning.
Answer: Pixels in BGR channel order
Convert to RGB if your model expects it.
Lesson: Preprocessing Consistency
Answer: Lower latency, offline operation and keeping images on site
Edge trades capacity for latency and privacy.
Lesson: Deploying Vision Models
Answer: Text descriptions of the classes, not labelled training images
Images are compared with text prompts.
Lesson: Vision Foundation Models
Answer: Counting on the device and discarding or blurring faces
Collect and keep only what you need.
Answer: Identical preprocessing in training and serving, verified with golden images
Consistency and realistic testing.