Non-Maximum Suppression (NMS)
Definition
NMS removes weaker, overlapping detection boxes in favor of stronger candidates.
A post-processing method that removes overlapping detection boxes with lower scores in favor of stronger candidates.
Simply explained
If a system detects the same object multiple times, redundant boxes are filtered out based on defined values.
Why it matters in practice
Helps realistically assess model quality, limits, and technical decisions.
Example from business
Multiple boxes around the same component are reduced to one box based on score and IoU.
Limits and pitfalls
- Choosing thresholds without validation or incorrectly merging nearby objects.
Origin & history
NMS has been a standard component since classical and neural object detectors.
Use in mid-sized companies
- Possible application: Multiple boxes around the same component are reduced to one box based on score and IoU.
Frequently asked questions
What does Non-Maximum Suppression mean?
A post-processing method that removes overlapping detection boxes with lower scores in favor of stronger candidates.
Which mistakes should be avoided?
Choosing thresholds without validation or incorrectly merging nearby objects.
Sources
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: