Skip to content
ml-recall

#non-maximum-suppression

4 questions from Object Detection

Timed test

4 questions · 4 min · score and explanations at the end

Take a test
Q 1 / 4Score 0 / 0

Question 1: Short answer

mediumAI-checkedObject Detection

Two people stand close together. A detector gives one box per person, with scores 0.95 and 0.80, and the boxes have an IoU of 0.6. Plain non-maximum suppression with threshold 0.5 would delete the 0.80 box. Linear Soft-NMS keeps it and instead multiplies its score by (1−IoU)(1 - \text{IoU}) whenever its IoU with the selected box is at least the threshold. What is its new score?

a number, 2 decimals