Home▸Machine Learning & Neural Networks▸YOLO Object Detection: Confidence Threshold & NMS (2D)

YOLO Object Detection: Confidence Threshold & NMS (2D)

A 2D companion to the YOLO detection lab: a field of candidate bounding boxes over a stylized scene, filtered live by a confidence-threshold slider and a real IoU (Intersection-over-Union) non-max-suppression pass, so you can watch raw predictions collapse into clean, labeled final detections.

Machine Learning & Neural Networks2DModerate60 FPS📱 Mobile-adapted⇄ 3D version

This 2D companion drives the same confidence-threshold and non-max-suppression logic as the 3D YOLO lab, but on a plain canvas view built for reading the box math directly: every candidate box carries a class and a confidence score, and a real IoU (Intersection-over-Union) computation decides which overlapping boxes of the same class get suppressed down to one final detection per object.

⚙ Under the hood

Confidence-threshold filtering plus a genuine per-class greedy non-max-suppression pass, driven by an IoU formula computed live from each box's coordinates — the same math that turns YOLO's dozens of overlapping candidates into clean final detections.

yoloobject detectionnon-max suppressionioucomputer vision

2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install