YOLO
In plain wordsA family of fast object-detection models that look at an image once and immediately output boxes and labels for everything they recognize, quick enough for live video.
A family of one-stage, real-time object detectorsobject detection, n.: Finding every object of interest in a picture or video frame, saying what each one is and drawing a box around it, for example all the cars, cyclists and pedestrians in a street scene. descended from 'You Only Look Once'. Region-based detectors such as R-CNNR-CNN, prop. n.: A family of object detectors that first pick out image regions likely to contain objects, then use a neural network to label each region and tighten its box. classified some 2,000 candidate regions and needed seconds per image; the original YOLO instead made every prediction in one forward pass of a single convolutional neural networkconvolutional neural network, n.: A kind of neural network built for images and other grid-like data. It slides small pattern detectors across the input, finding edges first, then shapes, then whole objects., reaching 45 frames per second at some cost in localization accuracy. It split the image into a 7 × 7 grid, and each cell output two bounding boxesbounding box, n.: The rectangle a vision system draws around an object it has found, stored as a handful of numbers giving the rectangle's position and size. with confidence scores and one set of class probabilities; the whole network was trained end to end. Successors added anchor boxesanchor box, n.: A ready-made box shape, placed at many spots across an image, that an object detector stretches and shifts to fit real objects instead of drawing boxes from scratch. (YOLOv2) and multi-scale prediction (YOLOv3), then anchor-free heads, separate classification and regression branches and heavy data augmentationdata augmentation, n.: Creating extra training examples by changing existing ones in ways that keep their meaning — a flipped, cropped or recolored photo of a cat still shows a cat — so the model learns more robustly.; recent releases also handle instance segmentationinstance segmentation, n.: Finding each separate object in an image and tracing its exact outline, so that two people standing together get two distinct shapes rather than one blob., pose estimationpose estimation, n.: Working out the position and orientation of something from images: where a person's joints are and how their body is posed, or exactly how an object or camera is placed and turned in space. and rotated boxes. YOLOv10 (2024) dropped non-maximum suppressionnon-maximum suppression, n.: A clean-up step in object detection: when the system draws several overlapping boxes around the same object, it keeps the most confident one and throws away the near-duplicates. to simplify deployment, and Ultralytics YOLO26 (released January 2026) offers an optional NMS-free prediction head. YOLO models are a default choice for detection on edge devices, drones, robots and video analytics.
We fine-tuned a small YOLO model on 3,000 labeled shelf photos and it runs at 60 fps on the store's camera box.
Origin
Introduced by Joseph Redmon, Santosh Divvala, Ross Girshick and Ali Farhadi (University of Washington, Allen Institute for AI, Facebook AI Research) in a June 2015 arXiv paper presented at CVPR 2016; the name puns on the slogan 'you only live once'. Redmon released YOLOv2 and YOLOv3, then left vision research in 2020 citing military and privacy uses; later versions came from other groups, notably the company Ultralytics.
Usage note
Version numbers are not one lineage: YOLOv4, v6, v7, v9, v10 and v12 come from different academic and industrial teams, and 'YOLO' now names a design style as much as a model. As of October 2026, Ultralytics had unveiled YOLO27 (September 2026) but not yet released it. Not to be confused with a YOLO runYOLO run, n.: Slang for launching a big, expensive training run without the usual careful small-scale tests first — changing many things at once and trusting experience and instinct that it will work. YOLO means 'you only live once'..
Related
- object detectionobject detection, n.: Finding every object of interest in a picture or video frame, saying what each one is and drawing a box around it, for example all the cars, cyclists and pedestrians in a street scene.
- R-CNNR-CNN, prop. n.: A family of object detectors that first pick out image regions likely to contain objects, then use a neural network to label each region and tighten its box.
- DETRDETR, prop. n.: An object detector built on the transformer, the design behind modern chatbots, that outputs a short list of objects in one go instead of sorting through thousands of guessed boxes.
- non-maximum suppressionnon-maximum suppression, n.: A clean-up step in object detection: when the system draws several overlapping boxes around the same object, it keeps the most confident one and throws away the near-duplicates.
- anchor boxanchor box, n.: A ready-made box shape, placed at many spots across an image, that an object detector stretches and shifts to fit real objects instead of drawing boxes from scratch.
- edge AIedge AI, n.: Doing AI's computing on or near the devices that collect the data, such as cameras, cars, factory machines, phones or local servers, instead of sending everything off to distant data centers.
- R-CNNR-CNN, prop. n.: A family of object detectors that first pick out image regions likely to contain objects, then use a neural network to label each region and tighten its box.
Sources
- Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. (2016). You Only Look Once: Unified, Real-Time Object Detection. CVPR 2016. arxiv.org/abs/1506.02640
- Sapkota, R. & Karkee, M. (2025). Ultralytics YOLO Evolution: An Overview of YOLO26, YOLO11, YOLOv8, and YOLOv5 Object Detectors for Computer Vision and Pattern Recognition. arXiv:2510.09653 (v4, September 2026, retitled to add YOLO27). arxiv.org/abs/2510.09653
- Ultralytics (2026). Key highlights from Ultralytics YOLO Vision 2026. Ultralytics blog. ultralytics.com/blog/key-highlights-from-ultralytics-yolo-vision-2026
Mentioned in
object detectionobject detection, n.: Finding every object of interest in a picture or video frame, saying what each one is and drawing a box around it, for example all the cars, cyclists and pedestrians in a street scene., bounding boxbounding box, n.: The rectangle a vision system draws around an object it has found, stored as a handful of numbers giving the rectangle's position and size., CVPRCVPR, prop. n.: The leading yearly conference on computer vision — teaching computers to understand images and video — and one of the most cited publication venues in any field of science., non-maximum suppressionnon-maximum suppression, n.: A clean-up step in object detection: when the system draws several overlapping boxes around the same object, it keeps the most confident one and throws away the near-duplicates., R-CNNR-CNN, prop. n.: A family of object detectors that first pick out image regions likely to contain objects, then use a neural network to label each region and tighten its box., anchor boxanchor box, n.: A ready-made box shape, placed at many spots across an image, that an object detector stretches and shifts to fit real objects instead of drawing boxes from scratch., DETRDETR, prop. n.: An object detector built on the transformer, the design behind modern chatbots, that outputs a short list of objects in one go instead of sorting through thousands of guessed boxes., YOLO modeYOLO mode, n.: Slang for letting an AI coding agent run any command without asking you first — fast and convenient, but risky, since nothing stops a harmful action.