YOLOv8: Extended Edition

📘 Summary

🚀 YOLOv8: Extended Edition is a fork of Ultralytics YOLOv8 repository. but with new features focuses on Automotive Vechicles 🚗.

🙋‍♂️ Who are we ? :

• We are a team from Egypt studied at Faculty of Engineering, Electronics and Communications Department in Mansoura University.

• These features were created to be used for our graduation project 🎓 (RTSD System) to make a system that can detect Roads Traffic Signs in real-time.

• We are using YOLOv8 as a base model and we are adding new features to it to make it more suitable for our use case.

If you are new to YOLOv8, please unpack the original YOLOv8 Readme below.

Original YOLOv8 Readme

English | 简体中文

Ultralytics YOLOv8 is a cutting-edge, state-of-the-art (SOTA) model that builds upon the success of previous YOLO versions and introduces new features and improvements to further boost performance and flexibility. YOLOv8 is designed to be fast, accurate, and easy to use, making it an excellent choice for a wide range of object detection and tracking, instance segmentation, image classification and pose estimation tasks.

We hope that the resources here will help you get the most out of YOLOv8. Please browse the YOLOv8 Docs for details, raise an issue on GitHub for support, and join our Discord community for questions and discussions!

To request an Enterprise License please complete the form at Ultralytics Licensing.

Documentation

See below for a quickstart installation and usage example, and see the YOLOv8 Docs for full documentation on training, validation, prediction and deployment.

Install

Pip install the ultralytics package including all requirements in a Python>=3.7 environment with PyTorch>=1.7.

pip install ultralytics

Usage

CLI

YOLOv8 may be used directly in the Command Line Interface (CLI) with a yolo command:

yolo predict model=yolov8n.pt source='https://ultralytics.com/images/bus.jpg'

yolo can be used for a variety of tasks and modes and accepts additional arguments, i.e. imgsz=640. See the YOLOv8 CLI Docs for examples.

Python

YOLOv8 may also be used directly in a Python environment, and accepts the same arguments as in the CLI example above:

from ultralytics import YOLO

# Load a model
model = YOLO("yolov8n.yaml")  # build a new model from scratch
model = YOLO("yolov8n.pt")  # load a pretrained model (recommended for training)

# Use the model
model.train(data="coco128.yaml", epochs=3)  # train the model
metrics = model.val()  # evaluate model performance on the validation set
results = model("https://ultralytics.com/images/bus.jpg")  # predict on an image
success = model.export(format="onnx")  # export the model to ONNX format

Models download automatically from the latest Ultralytics release. See YOLOv8 Python Docs for more examples.

Models

All YOLOv8 pretrained models are available here. Detect, Segment and Pose models are pretrained on the COCO dataset, while Classify models are pretrained on the ImageNet dataset.

Models download automatically from the latest Ultralytics release on first use.

Detection

See Detection Docs for usage examples with these models.

Model	size (pixels)	mAPval 50-95	Speed CPU ONNX (ms)	Speed A100 TensorRT (ms)	params (M)	FLOPs (B)
YOLOv8n	640	37.3	80.4	0.99	3.2	8.7
YOLOv8s	640	44.9	128.4	1.20	11.2	28.6
YOLOv8m	640	50.2	234.7	1.83	25.9	78.9
YOLOv8l	640	52.9	375.2	2.39	43.7	165.2
YOLOv8x	640	53.9	479.1	3.53	68.2	257.8

• mAP^val values are for single-model single-scale on COCO val2017 dataset.
  Reproduce by yolo val detect data=coco.yaml device=0
• Speed averaged over COCO val images using an Amazon EC2 P4d instance.
  Reproduce by yolo val detect data=coco128.yaml batch=1 device=0|cpu

Segmentation

See Segmentation Docs for usage examples with these models.

Model	size (pixels)	mAPbox 50-95	mAPmask 50-95	Speed CPU ONNX (ms)	Speed A100 TensorRT (ms)	params (M)	FLOPs (B)
YOLOv8n-seg	640	36.7	30.5	96.1	1.21	3.4	12.6
YOLOv8s-seg	640	44.6	36.8	155.7	1.47	11.8	42.6
YOLOv8m-seg	640	49.9	40.8	317.0	2.18	27.3	110.2
YOLOv8l-seg	640	52.3	42.6	572.4	2.79	46.0	220.5
YOLOv8x-seg	640	53.4	43.4	712.1	4.02	71.8	344.1

• mAP^val values are for single-model single-scale on COCO val2017 dataset.
  Reproduce by yolo val segment data=coco.yaml device=0
• Speed averaged over COCO val images using an Amazon EC2 P4d instance.
  Reproduce by yolo val segment data=coco128-seg.yaml batch=1 device=0|cpu

Classification

See Classification Docs for usage examples with these models.

Model	size (pixels)	acc top1	acc top5	Speed CPU ONNX (ms)	Speed A100 TensorRT (ms)	params (M)	FLOPs (B) at 640
YOLOv8n-cls	224	66.6	87.0	12.9	0.31	2.7	4.3
YOLOv8s-cls	224	72.3	91.1	23.4	0.35	6.4	13.5
YOLOv8m-cls	224	76.4	93.2	85.4	0.62	17.0	42.7
YOLOv8l-cls	224	78.0	94.1	163.0	0.87	37.5	99.7
YOLOv8x-cls	224	78.4	94.3	232.0	1.01	57.4	154.8

• acc values are model accuracies on the ImageNet dataset validation set.
  Reproduce by yolo val classify data=path/to/ImageNet device=0
• Speed averaged over ImageNet val images using an Amazon EC2 P4d instance.
  Reproduce by yolo val classify data=path/to/ImageNet batch=1 device=0|cpu

Pose

See Pose Docs for usage examples with these models.

Model	size (pixels)	mAPpose 50-95	mAPpose 50	Speed CPU ONNX (ms)	Speed A100 TensorRT (ms)	params (M)	FLOPs (B)
YOLOv8n-pose	640	50.4	80.1	131.8	1.18	3.3	9.2
YOLOv8s-pose	640	60.0	86.2	233.2	1.42	11.6	30.2
YOLOv8m-pose	640	65.0	88.8	456.3	2.00	26.4	81.0
YOLOv8l-pose	640	67.6	90.0	784.5	2.59	44.4	168.6
YOLOv8x-pose	640	69.2	90.2	1607.1	3.73	69.4	263.2
YOLOv8x-pose-p6	1280	71.6	91.2	4088.7	10.04	99.1	1066.4

• mAP^val values are for single-model single-scale on COCO Keypoints val2017 dataset.
  Reproduce by yolo val pose data=coco-pose.yaml device=0
• Speed averaged over COCO val images using an Amazon EC2 P4d instance.
  Reproduce by yolo val pose data=coco8-pose.yaml batch=1 device=0|cpu

Integrations

Roboflow	ClearML ⭐ NEW	Comet ⭐ NEW	Neural Magic ⭐ NEW
Label and export your custom datasets directly to YOLOv8 for training with Roboflow	Automatically track, visualize and even remotely train YOLOv8 using ClearML (open-source!)	Free forever, Comet lets you save YOLOv8 models, resume training, and interactively visualize and debug predictions	Run YOLOv8 inference up to 6x faster with Neural Magic DeepSparse

Ultralytics HUB

Experience seamless AI with Ultralytics HUB ⭐, the all-in-one solution for data visualization, YOLOv5 and YOLOv8 🚀 model training and deployment, without any coding. Transform images into actionable insights and bring your AI visions to life with ease using our cutting-edge platform and user-friendly Ultralytics App. Start your journey for Free now!

Contribute

We love your input! YOLOv5 and YOLOv8 would not be possible without help from our community. Please see our Contributing Guide to get started, and fill out our Survey to send us feedback on your experience. Thank you 🙏 to all our contributors!

License

YOLOv8 is available under two different licenses:

• AGPL-3.0 License: See LICENSE file for details.
• Enterprise License: Provides greater flexibility for commercial product development without the open-source requirements of AGPL-3.0. Typical use cases are embedding Ultralytics software and AI models in commercial products and applications. Request an Enterprise License at Ultralytics Licensing.

Contact

For YOLOv8 bug reports and feature requests please visit GitHub Issues, and join our Discord community for questions and discussions!

⚡ What's New!

We added new features to YOLOv8, including:

• 🌚 Night Vision:
  YOLOv8 can now see in the dark! We added support for night vision cameras, which can be used to detect objects in low-light conditions.

• 🛣 Lane Line Detection:
  YOLOv8 can now detect lane lines on the road! This feature is useful for self-driving cars and other autonomous vehicles.

• 🔌 SPI output:
  YOLOv8 can now output data over SPI, which is useful for connecting to other devices such as Arduino boards, Raspberry Pis or ESP32s.

📚 Documentation

We will talk about each of these features in more detail below.

---

🌚 Night Vision

Unpack Night Vision

Note:
     This feature is available for detect mode.
     and val mode, only with device=cpu.

How does it work ⚙?

Gamma correction is pretty simple, it's just a non-linear transformation of the input image.
It is used to adjust the overall brightness of the image.
The formula for gamma correction is as follows:

where Image_out is the output normalized image, Image_in is the input normalized image, and Gamma is the gamma value (power).

Note:
     1- The value of gamma is between 0 and 1 (closer to 0 brighter, closer to 1 darker).
     2- The value of both Image_out and Image_in pixels are between 0 and 1.

To make it more clear, let's take a look at the following example:

Let's assume we have an image with only one pixel, and its value is 0.8 (normalized value).
Now, let's apply gamma correction as follows:

Image_out = Image_in ^ gamma

when gamma = 0, then Image_out = 1
when gamma = 0.25, then Image_out = 0.94
when gamma = 0.5, then Image_out = 0.89
when gamma = 0.75, then Image_out = 0.84
when gamma = 1, then Image_out = 0.8

let's plot the results:

Note:
     Decreasing the value of gamma will make the image brighter.

How to use it 🚄?

There are some arguments added for Night Vision Feature :

1. night_vision
2. image_gamma (Optional)

   • if integer or float number :

     • No need to pass any other parameters.

   • if 'auto', then you can pass the following parameters (all of them have values between 0 and 1):

     • 🟢 min_normalized_intensity (Optional)
     • 🟢 min_gamma (Optional)
     • 🟢 max_gamma (Optional)

these parameters are described as below:

🔵 night_vision parameter :

(default value is False)

If you use camera in a dark environment, then you may get poor results.
We have made a preprocess feature to enhance image brightness, it actually may help you get better results in dark environments.

You can use Night Vision mode by adding parameter night_vision as follows:

Using CLI:

yolo detect predict model=path/to/best.pt source='https://ultralytics.com/images/bus.jpg' night_vision

and it has 3 modes:

1- Default mode:

By not passing parameter night_vision, it will use the original image as it is.

2- Night Vision mode (Only apply on input image):

night_vision

Or

night_vision=true

It will apply night vision on input image, pass it to the model and get the result based on night-processed image, But the shown/saved image will be the original image without night filter.

3- Night Vision mode (Apply on both input image and shown/saved image):

night_vision=show

It will apply night vision on input image, pass it to the model and get the result based on night-processed image, The shown/saved image will be also night-processed image with night filter applied.

An example to differentiate between Night Vision modes and their (Saved/Shown) results:

🔵 image_gamma parameter :

(default value is auto)

image_gamma parameter has 2 modes:

1- Fixed gamma value:

If you want to use fixed gamma value (from 0 to 1), you can pass image_gamma parameter as follows:
     Note 1: 1 means no change in image brightness, 0 means white image.
     Note 2: Lower gamma value means more lightening applied to image (brighter image).

Using CLI:

yolo detect predict model=path/to/best.pt source='https://ultralytics.com/images/bus.jpg' night_vision=show image_gamma=0.8

It will use fixed gamma value of 0.8

2- Auto gamma value:

If you want to use auto gamma value, you can pass image_gamma parameter as follows:

Using CLI:

yolo detect predict model=path/to/best.pt source='https://ultralytics.com/images/bus.jpg' night_vision=show image_gamma=auto

It will test normalized image intensity (its value from 0 to 1), and its value describe image brightness (closer to 0 means dark image, closer to 1 means bright image).

if it is less than min_normalized_intensity (default value is 0.25) meaning that image is dark, then it will use scaled gamma value (from min_gamma to max_gamma) - based on image intensity - to enhance image brightness.
     Note: min_gamma and max_gamma have default values of 0.6 and 1.0 respectively.

if it is greater than min_normalized_intensity (default value is 0.25) meaning that image is bright, then it will use fixed gamma value of 1 (no change in image brightness).

For full customized gamma value, you can pass min_normalized_intensity, min_gamma and max_gamma as follows:

Using CLI:

yolo detect predict model=path/to/best.pt source='https://ultralytics.com/images/bus.jpg' night_vision=show image_gamma=auto min_normalized_intensity=0.5 min_gamma=0.8 max_gamma=1.0

An example to show effect of gamma (image_gamma) on image brightness :

---

🛣 Lane Line Detection

Unpack Lane Line Detection

Lane Line Detection is a feature that can be used to detect lane lines in images and videos.

How does it work ⚙?

Lane Line Detection is a feature that can be used to detect lane lines in images and videos.

The algorithm is based on the following steps:

1. Detecting edges using Canny Edge Detection.
2. Applying a mask to the image to remove unnecessary parts.
3. Applying Hough Transform to detect lines.
4. Filtering the detected lines to get the left and right lane lines.
5. Drawing the detected lane lines on the image.

How to use it 🚄?

There are some arguments added for Lane Line Detection Feature :

1. lane_detection:
   • if lane_detection is True, then it will apply lane line detection on the input image.
   • if lane_detection is False, then it will not apply lane line detection on the input image.

   Note:
        Default value is False.

2. Optional parameters:
   • CANNY_THRESHOLD_1: The first threshold for the hysteresis procedure in Canny Edge Detection. (default value is 50)
   • CANNY_THRESHOLD_2: The second threshold for the hysteresis procedure in Canny Edge Detection. (default value is 150)
   • MIN_VOTES: The minimum number of votes (intersections in Hough grid cell). (default value is 100)
   • MIN_LINE_LEN: The minimum number of pixels making up a line. (default value is 50)
   • MAX_LINE_GAP: The maximum gap between two points to be considered in the same line. (default value is 10)
   • lane_line_color: The color of the detected lane lines. (default value is [243, 105, 14])
   • lane_line_thickness: The thickness of the detected lane lines. (default value is 12)

You can use Lane Line Detection by adding parameter lane_detection as follows:

Using CLI:

yolo detect predict model=path/to/best.pt source='https://ultralytics.com/images/bus.jpg' lane_detection

and it has 2 modes:

1- Default mode:

By not passing parameter lane_detection, it will not apply lane line detection on the input image.

2- Lane Line Detection mode:

lane_detection

Or

lane_detection=true

It will apply lane line detection on input image, pass it to the model and get the result based on lane line detected image, The shown/saved image will have lane line detected.

An example to differentiate between Lane Line Detection modes and their (Saved/Shown) results:

---

🔌 SPI output

Unpack SPI output

SPI output is a feature that can be used to output data over SPI, which is useful for connecting to other devices such as Arduino boards, Raspberry Pis or ESP32s.

How does it work ⚙?

SPI stands for Serial Peripheral Interface. It is a synchronous serial communication interface specification used for short-distance communication, primarily in embedded systems.
SPI devices communicate in full duplex mode using a master-slave architecture with a single master. The master device originates the frame for reading and writing. Multiple slave devices are supported through selection with individual slave select (SS) lines.

How to use it 🚄?

There are some arguments added for SPI output Feature :

1. spi:
   • if spi is True, then it will output data over SPI.
   • if spi is False, then it will not output data over SPI.

   Note:
        Default value is False.

2. Optional parameters:
   • spi_mode: The SPI mode to use. (default value is 3)
   • spi_speed: The SPI speed to use. (default value is 2000000)
   • spi_sleep: The SPI delay to use. (default value is 0)
   • spi_port: The SPI port to use. (default value is 0)
   • spi_device: The SPI device to use. (default value is 0)

You can use SPI output by adding parameter spi as follows:

Using CLI:

yolo detect predict model=path/to/best.pt source='https://ultralytics.com/images/bus.jpg' spi

and it has 2 modes:

1- Default mode:

By not passing parameter spi, it will not output data over SPI.

2- SPI output mode:

spi

Or

spi=true

It will output data over SPI.

---

🙏 Thanks for reading!

We hope our new features will help you in your projects.