Decoder. plain-English AI glossary

Object Detection

● Core

Identifying and locating multiple objects in an image—boxing where each is and labeling what it is.

Think of it like

Circling every person in a crowd photo and writing their name.

Example

YOLO detects "person" at (100, 50, 300, 400), "car" at (450, 200, 600, 350) in one forward pass.

How it actually works

Classification says "this image has dogs"; detection says "dog at top-left, dog at bottom-right." Requires bounding box regression (predict coordinates) plus classification. YOLO is fast one-stage detector; Faster R-CNN is two-stage (region proposal then classification). Training requires bounding box annotations, expensive to collect.

For product teams

Enables autonomous driving, security monitoring, content moderation (what objects appear in a scene?).

For engineers

Network predicts class probabilities and bounding boxes for each anchor; NMS (non-maximum suppression) removes duplicate detections.

Related

Read anything AI without the jargon

Look up any term in plain English, or save terms as you read with the free Chrome extension.

Open DecoderAdd to Chrome