Detection Is the Entry Gate | Visual AI Learning Guide

That first stage is doing more work than it looks like. Here is a normal scene: several faces, a table, a window, a lot of pixels that have nothing to do with anyone's identity.

0:00 / 0:00

Auto-play

Detection is the stage that decides what the rest of the system is even allowed to look at. Given a full scene containing several faces and background objects, a detector scans the image and produces a bounding box around each face it finds. Only the pixels inside that box are passed forward; the background, the furniture, and every other object are discarded. This is why detection must come first: the embedding stage expects a cropped face region, not a whole photograph, and a crop that is off-center or too loose degrades everything downstream.