Glossary

Face detection

Locating faces in an image and returning a box and confidence for each.

Definition

Face detection is the computer-vision task of finding where faces are in an image or video and returning, for each one, a bounding box and a confidence score. It answers "where are the faces?", not "whose face is this?" — that second question is face recognition, a separate and more sensitive task.

01

Why it matters

Detection is the first step of anonymization: you cannot blur a face until you know where it is. Because a detector has a miss rate — on profiles, reflections, small or motion-blurred faces — a missed detection is a leaked identity, so recall and human review matter more than a single accuracy figure.

02

How Firsthand handles it

Firsthand runs face detection on every frame at ingest, tunes the confidence threshold to the recall target in your spec, tracks each box across frames, and confirms the result with a human reviewer before delivery.