‍

YOLO (You Only Look Once) is one of the most popular object detection models. It is known for its speed and accuracy. It processes images in real time, making it useful for applications like autonomous driving, surveillance, and robotics.

Here we will cover:

• What is YOLO?
• How does YOLO work?
• Evolution of YOLO: From v1 to v8
• How to implement YOLO
• Usecases and applications

By the end, you'll understand how YOLO works, its strengths and trade-offs, and how to use it for various object detection tasks.

What is YOLO for Object Detection?

The YOLO is a real-time object detection model which can process an entire image in a single pass. Introduced by Joseph Redmon et al. in 2015, YOLO reframed object detection as a single end-to-end regression problem. It directly maps image pixels to bounding box co-ordinates and class probabilities. This design made YOLO significantly faster than previous approaches.

Why is YOLO Revolutionary?

Previously the popular object detection models like Fast R-CNN used a two stage approach. They would first generate region proposals and then classify them, which made object detection complex and cannot be processed in real-time. 

YOLO used a single convolutional neural network (CNN) and eliminated the region proposal step. This was revolutionary as it made the process simple and enabled real-time detection with competitive accuracy.

The first version of YOLO had lower localization accuracy compared to two-stage methods, but later versions (YOLOv2, v3, etc.) closed this gap. The ability to process at 30+ FPS with high mean Average Precision (mAP) made YOLO practical for real-time applications like video analysis, drone vision, and mobile object detection.

One-Stage vs. Two-Stage Detectors

Object detection models are typically categorized into: two-stage and one-stage detectors. The key difference lies in how they process an image to detect objects.

Two-Stage Detectors: High Accuracy, Slower Speed

Two-stage detectors, like Faster R-CNN, break object detection into two separate steps:

1. Region Proposal: A Region Proposal Network (RPN) scans the image and suggests potential object locations.
   
   
2. Classification & Refinement: Each proposed region is classified and refined to improve accuracy.

This method is highly accurate as the deep learning model focuses on likely object regions and then classifies the potential objects. However, this also adds computations and makes it slower, typically achieving 5-7 FPS on a high-end GPU.

One-Stage Detectors: Faster, Real-Time Performance

One-stage detectors, like YOLO and Single Shot Detection or SSD, skip the region proposal step and predict bounding boxes and class labels in a single network pass. This direct approach makes them significantly faster.

‍

YOLO was one of the first one-stage detectors to achieve high accuracy, outperforming earlier single-shot models like SSD. 

YOLO vs. Other Object Detection Algorithms

While YOLO dominates in speed, it’s useful to understand how it compares with other detection frameworks.

YOLO vs. Faster R-CNN

Faster R-CNN uses a Region Proposal Network (RPN) to generate ~300 object regions before classification. It achieves high accuracy but runs at roughly 5-7 FPS with a ResNet-101 backbone. YOLOv3, in contrast, runs at 20-45 FPS with slightly lower accuracy. While two-stage models historically had better localization for small objects, YOLOv7 has surpassed many two-stage models in accuracy.

YOLO vs. SSD

Single shot multibox detector or SSD also uses a single stage approach like YOLO with multi-scale feature maps and anchor boxes. But with YOLO v4 there was a significant improvement in the accuracy over SSD.

YOLO vs. RetinaNet

The RetinaNet used Focal Loss to handle class imbalance in classification and achieved accuracy comparable to two-stage detectors. It improved the detection of small objects but at the cost of speed. Later YOLO versions (v4, v5) outperformed RetinaNet in both speed and accuracy, making YOLO the better choice for real-time tasks.

YOLO vs. EfficientDet

EfficientDet has a pre trained model as a backbone followed by BiFPN as a feature network. This improved the accuracy, but at lower speeds. EfficientDet-D4 matched YOLOv4’s accuracy but ran at ~8-11 FPS, while YOLOv4 achieved 62 FPS. Even EfficientDet-D7X, the most accurate variant, was slower than YOLOv7. YOLOv7 outperformed it in accuracy as well.

‍

‍

How YOLO Object Detection Works (Single-Shot Detection)

How YOLO Object Detection Works:

‍

Here are some of the key components involved in the YOLO object detection algorithm:

Grid Division and Object Localization

YOLO first divides the input image into an S x S grid. Now each grid cell is used to detect objects that falls within it.

Class and Bounding Box Prediction

Each bounding box is defined by its center coordinates, width, height, and a confidence score that indicates the likelihood of an object being present. The model also assigns class probabilities to each grid cell, allowing it to identify different objects in a single inference step.

Non-Maximum Suppression (NMS)

The algorithm often predicts multiple overlapping boxes for the same object. To eliminate duplicates, the Non Maximum Algorithm filters out boxes with lower confidence. This ensures that the detections are not redundant.

Confidence Score

This score represents the probability that a bounding box contains an object and how well the predicted box fits the object. The confidence score here is different from the object confidence score assigned earlier to each bounding box.

Multi-Scale Detection in Later Versions

The earlier versions of YOLO struggled with small object detection tasks. YOLOv2 introduced multi-scale feature extraction and used FPN to detect objects at different resolutions.

YOLO's single-shot architecture enables real-time performance, even on local machines. Its balance of speed, accuracy, and accessibility has made it widely adopted in various applications.

Evolution of YOLO Models: From v1 to v12

Since its introduction in 2015, YOLO has evolved through multiple versions. It improved on YOLO architecture in each iteration to enhance accuracy, efficiency, and adaptability over the years. Here is an overview of each iteration:

YOLOv1 (2015): The Original YOLO

YOLOv1 was the first to unify object detection into a single neural network. It used a 24 layer CNN similar to GoogLeNet and predicted two bounding boxes per cell across 20 classes.

Key Features

• Single-stage object detection using a 7×7 grid.
• Introduced grid-cell responsibility for object localization.
• Predicts bounding boxes and class probabilities in a single forward pass.
• Realtime performance: process images at 45 frames per second.

Performance

YOLOv1 achieved 63.4% mAP on PASCAL VOC 2007 at 45 FPS on a GPU. However, it had lower accuracy than region-based methods like Faster R-CNN, particularly for small and overlapping objects.

Impact

YOLOv1 proved real-time object detection was feasible on a single GPU.

YOLOv2 (2016): Better and Faster

YOLOv2, also called YOLO9000, improved upon v1 by incorporating anchor boxes, a new backbone (Darknet-19), and batch normalization, allowing detection of over 9000 object categories via joint training on ImageNet.

Key Features

• Introduced anchor boxes for better localization.
• Also introduced batch normalization and high-resolution classification.
• Used multi-scale training for better generalization across different object sizes.
• High-resolution classifiers improved detection accuracy.

Performance

YOLOv2 runs at 67 FPS with a 78.6 mAP on VOC 2007, and 21.6% AP on VOVO at 40 FPS. It surpasses YOLOv1 in both speed and accuracy.

Impact

YOLOv2 bridged the performance gap with state-of-the-art detectors while maintaining real-time speed, making it practical for industry applications.

YOLOv3 (2018): Multi-Scale Predictions

YOLOv3 introduced a deeper backbone (Darknet-53) compared to Darket-19 used in YOLOv2. It uses the backbone with residual connections and a feature pyramid network (FPN) for multi-scale object detection tasks.

Key Features

• Multi-scale predictions at 3 different resolutions (13×13, 26×26, 52×52).
• Uses logistic classifiers for predicting object classes instead of softmax, allowing for multi-label classification.
• Uses anchor boxes with different scales and aspect ratios to better match the size and shape of the objects being detected.

Performance

YOLOv3 achieved 3o FPS with 33% mAP while significantly improving detection accuracy, especially for small objects. It was a strong competitor to Faster R-CNN, SSD and RetinaNet and is 3-4 times faster. Hence it was preferred for practical applications.

Impact

YOLOv3 became a widely used real-time detector, balancing speed and accuracy. However, in 2020, Redmon ceased research on YOLO, leaving further development to the community.

YOLOv4 (2020): Community-Driven Enhancements

Developed by Bochkovskiy, Wang, and Liao, YOLOv4 improved both speed and accuracy using CSPDarknet-53 as a backbone and numerous architectural optimizations.

Key Features

• Introduced CSPNet for reduced computation.
• Used Mish activation for better feature learning.
• Added Mosaic data augmentation and Self-Adversarial Training (SAT).
• Improved loss functions (CIOU loss) and regularization techniques.

Performance

YOLOv4 reached 62 FPS on a Tesla V100, offering a superior speed-accuracy balance. It surpassed YOLOv3 in both mAP and efficiency.

Impact

YOLOv4 established itself as the top choice for real-time object detection in 2020, gaining widespread adoption in research and industry.

YOLOv5 (2020): PyTorch Implementation

YOLOv5, released by Ultralytics, was the first major YOLO version implemented in PyTorch. While not an official research paper, it became extremely popular due to its ease of use and modular framework.

Key Features

• PyTorch implementation for easy training and deployment.
• Smaller and faster deep learning models (YOLOv5s, YOLOv5m, etc.).
• Augmentation techniques (Mosaic, MixUp, etc.) and improved anchor selection.

Performance

It was faster compared to YOLOv4 in terms of training and inference. The accuracy was competitive across benchmarks.

‍

Impact

YOLOv5 became widely adopted in computer vision applications because of its ease of use and performance. Also, it was optimized for mobile deployment.

YOLOv6 (2022): Industry-Focused Efficiency

YOLOv6 was developed by Meituan and optimized for industrial applications. It focused on efficiency and introduced an anchor-free architecture to improve detection accuracy and speed.

Key Features

• Introduces an anchor-free design and simplifies the training process, and enhances speed.
• Uses RepVGG-based structures for optimized feature extraction.
• Better optimization techniques with Knowledge Distillation and Quantization to improve performance.

Performance

It achieves a higher FPS than YOLOv5 while maintaining competitive accuracy.

Impact

It was optimized for edge deployment in industrial scenarios. Hence, it was widely used in manufacturing and automation due to its low latency and high speed inference capabilities.

YOLOv7 (2022)

YOLOv7 was developed by WongKinYiu and AlexeyAB as an independent research effort, focusing on balancing speed and accuracy. It introduced efficient reparameterization techniques.

Key Features

• Introduced Extended Efficient Layer Aggregation Networks (E-ELAN) which improves gradient flow for better training.
• Enhances inference speed by using reparameterized convolutions.
• Has multiple model variants applicable for various applications.

Performance

It is faster and more accurate than YOLOv6 and YOLOv5. It achieved higher mAP at lower latency compared to previous YOLO versions.

Impact

It was used in real-time video analytics and robotics due to its high accuracy and efficiency.

YOLOv8 (2023)

YOLOv8 refined previous improvements with a more flexible architecture, optimized for various real-world applications.

Key Features

• Used a new backbone and head which optimized the model for high accuracy and fast inference.
• Supports instance segmentation and object tracking.
• Further improved feature extraction and detection accuracy.
• Optimizes anchor boxes for better performance with AutoAnchor mechanism.
• Designed for better adaptability across different deployment environments.

Performance

YOLOv8 achieved higher accuracy and better generalization while keeping real-time performance intact. It remains one of the most widely used single-shot object detection models today.

Impact

It became popular version of YOLO due to its ease of use and high accuracy. It was commonly used in autonomous vehicles, surveillance and retail analytics.

YOLOv9 (2024)

YOLOv9 introduced a hybrid anchor-free detection approach, optimizing speed and accuracy. It refined feature aggregation and backbone efficiency for improved small-object detection in real-time applications.

Key Features

• Used generalized efficient layer aggregation network (GELAN) architecture to improve feature extraction and gradient flow.
• Optimizes the path aggregation network for better feature fusion across scales.
• Explores the functionality of multi-level auxiliary information, using different feature pyramids for varied tasks in object detection.
  

Performance

YOLOv9 demonstrates improved mAP over YOLOv8, with reduced latency, making it suitable for applications requiring swift and accurate object detection. YOLOv9 is also capable of performing object detection, segmentation, and classification tasks.

Impact

Better application in various industries because of its use.

YOLOv10 (2024)

YOLOv10 also known as YOLOE integrated transformer-based feature extraction, boosting performance in complex real-world scenarios. It improved generalization across diverse datasets while reducing computational overhead.

Key Features

• Introduces a dual-label assignment system to improve the model's capability to detect and classify objects in real-time.
• Introduces different prompt mechanisms like texts, visual inputs and prompt free paradigm.
• Lightweight classification head is used to balance accuracy and computational efficiency.
• Uses partial self attention or PSA modules to improve performance without significantly increasing computational cost.

Performance

YOLOv10 variants exhibit significant improvements over previous versions, achieving up to 54.4% APval with reduced latency. It is optimized for real-time edge computing applications, processing images at up to 1000fps.

Impact

YOLOv10 offers a range of model zises to accommodate different computational resources and accuracy needs. This efficiency driven design has set new benchmarks for real-time object detection, making it ideal for applications in resource constrained environments.

YOLOv11 

YOLOv11 shifts from purely CNN based architecture to  transformer based backbone. It introduces dynamic head design for improving accuracy with fewer parameters. It supports tasks such as object detection, segmentation, classification, keypoint detection, and oriented bounding box detection.

Key Features

• The dynamic head design adapts based on image complexity and optimizes resource allocation.
• Eliminates the need for Non Maximum Suppression and reduces inference time.
• Uses dual label assignment to improve detection in overlapping and densely packed objects.

Performance

YOLOv11 outperforms previous versions in speed and accuracy on COCO dataset. It processes at 60 FPS with a mean Average Precision (mAP) of 61.5%, and fewer parameters, making it suitable for a range of applications.

Impact

It utilizes a better neck and backbone architecture, enhancing feature extraction capabilities for more precise object detection. YOLOv11 also expanded object detection use cases, particularly for dense scenes and complex environments.

YOLOv12 (2025)

YOLOv12 integrates attention mechanisms into the YOLO framework. This design combines CNN speed with transformer-based enhancements. 

Key Features

• Uses Area Attention module which divides the featuremap into segments to preserve a large receptive field and reduces computational complexity.
• Addresses optimization challenges introduced by attention mechanisms with residual efficient layer aggregation network (R-ELAN).
• Introduces flash attention into the network to optimize memory.

Performance

It shows a 25% improvement in detection accuracy in poor lighting. The multiple object tracking enhances object tracking in motion heavy scenarios.

Impact

Sets a new benchmark in object detection with improved speed and accuracy. This makes YOLOv12 particularly effective in applications such as autonomous driving, security surveillance, and industrial automation.

YOLO Series - Comparison

Take a look at this comparison table.

‍

‍

Tools and Frameworks for Implementing YOLO

To train, fine-tune or even infer a YOLO model, you will need the right tools. Here are the libraries, frameworks and deployment solutions you will need to train, fine-tune or even infer your YOLO: 

Frameworks

PyTorch

It is a fan favorite for a good reason. It is flexible, easy to debug and has great support for GPU acceleration. Most modern YOLO versions are built on PyTorch as well.

You can install it with:

pip install torch torchvision

TersorFlow and Keras

YOLOv3 and YOLOv4 have TensorFlow implementations, and TensorFlow Lite (TFLite) makes it easy to deploy on mobile devices.

To get started:

pip install tensorflow

Darknet

Darknet is where YOLO started. It’s a C-based framework built for speed. While newer YOLO versions have moved to PyTorch, Darknet still supports YOLOv1 to YOLOv4 and YOLOv7 versions.

You can start with:

git clone https://github.com/pjreddie/darknet
cd darknet
make

Model Training and Deployment Tools

Ultralytics

It is a PyTorch-based implementation that simplifies training, fine-tuning, and deployment of YOLO models.

pip install ultralytics

Now to run inference on an image, run:

from ultralytics import YOLO

model = YOLO("yolov8n.pt")
results = model("image.jpg")

MMDetection

A PyTorch-based object detection framework developed by OpenMMLab. If you want more customization, then it is a solid choice. It’s modular and great for large-scale training.

pip install mmdet

ONNX Runtime

If you need to run YOLO models on a different hardware, then ONNX Runtime helps you convert your model so it can run on CPUs, GPUs or even on AI chips.

import onnxruntime as ort

session = ort.InferenceSession("yolov8.onnx")

Optimization Tools

• NVIDIA TensorRT: If you’re using an NVIDIA GPU, TensorRT is a must. It significantly reduces YOLO’s latency.
  
  
• OpenVINO: When using Intel’s hardware, OpenVINO helps in optimizing the model for CPUs and edge applications. It also reduces latency and power consumption.

Deploying YOLO Models

• Flask and FastAPI: Web frameworks for deploying YOLO models as APIs.
  
  
• NVIDIA Jetson Nano and Raspberry Pi can run optimized YOLO models on edge devices for real-time applications.

When not to use YOLO

• High-precision tasks: YOLO trades accuracy for speed; models like Faster R-CNN work better for detailed detections.
• Small or overlapping objects: Struggles with tiny objects in dense scenes.
• Complex relationships: Not ideal for tasks needing multi-stage processing or object tracking.
• Limited hardware: Requires GPUs for real-time performance; MobileNet SSD is better for low-power devices.

Use Cases and Applications of YOLO

YOLO’s combination of speed and accuracy has led to its adoption in a wide range of fields. Here are some prominent use cases:

• Autonomous Vehicles: Detects pedestrians, traffic signs, and other vehicles in real-time.
• Surveillance & Security: Enables real-time threat detection in CCTV footage.
• Retail & Inventory Management: Tracks products and automates checkout systems.
• Healthcare & Medical Imaging: Assists in detecting abnormalities in X-rays and MRIs.
  Robotics:  Helps robots recognize and interact with objects in dynamic environments.
• Sports Analytics: Tracks player movements and ball trajectories in live games.
• Augmented Reality (AR): Enhances AR applications by detecting objects for interactive overlays.

Conclusion

YOLO (You Only Look Once) has come a long way, evolving into one of the fastest and most efficient single shot object detection models out there. From YOLOv1 to the latest YOLOv12, each version has pushed the boundaries of speed, accuracy, and efficiency.

While YOLO remains a top choice for real-time vision tasks, selecting the right version ensures optimal performance for specific use cases.