YOLO (You Only Look Once): A brief introduction

Exploring the fundamentals of Object Detection

Image created by the author with DALL E - 3

Table of contents

1. How YOLO works — general notion
2. Image annotation
3. YOLO architecture
4. YOLO training process
5. Non-maximum suppression
6. Intersection Over Union (IoU)
7. Practical example to understand better YOLO
8. Loss function

You Only Look Once (YOLO) is a groundbreaking type of Convolutional Neural Network in the field of object detection.

Unlike traditional object detection systems that process an image in multiple steps, YOLO simplifies this task into a single step, making it incredibly fast and efficient.

The key idea:

YOLO frames object detection as a single regression problem.

Object detection typically involves two key components:

1. Identifying where objects are in an image — localization.
2. Determining what those objects are — classification.