Skip to content

YOLOv8-3D is a LowCode, Simple 2D and 3D Bounding Box Object Detection and Tracking , Python 3.10

Notifications You must be signed in to change notification settings

bharath5673/YOLOv8-3D

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

YOLOv8-3D (tracker)

Description:

YOLOv8-3D is a lightweight and user-friendly library designed for efficient 2D and 3D bounding box object detection in Advanced Driver Assistance Systems (ADAS). With its intuitive API and comprehensive features, EasyADAS makes it straightforward to integrate object detection capabilities into your ADAS projects.

YOLOv8-3D

This API supports for easy understanding and integrate 3D perception, systems can make more informed decisions and operate effectively in complex, real-world environments.

Key Features:

  • Simplified API: YOLOv8-3D provides a straightforward API that allows you to quickly implement object detection in your ADAS system.
  • 2D and 3D Bounding Boxes: Detect both 2D and 3D bounding boxes for accurate spatial understanding of objects.
  • Efficient Processing: Leverage optimized algorithms for fast and reliable object detection performance.
  • Flexible Integration: EasyADAS is designed to seamlessly integrate with existing ADAS systems and frameworks.
  • Comprehensive Documentation: Extensive documentation and examples ensure that you can get started quickly and easily.
  • Scalable: Scale your detection capabilities by leveraging EasyADAS in a variety of scenarios and environments.
  • supports all best and and state of the art CNN models and easy training setup.
  • added augmentations for 3d .
  • api supports Resnets, VGG nets, mobilenets, Efficientnets , MOT-Tracker

special features

augmentations for better training, automated backup training and results plot

Get started

Prerequisites

Usage

Installation

  1. Clone this repository.
  2. Install the required dependencies

Run

for training

conda create -n test1 python=3.10 -y
conda activate test1
pip install tensorflow

For more detailed tensorflow gpu installation instructions and options, refer to this documentation.

####### select model on train.py  ########
# select_model = 'resnet50'
# select_model ='resnet101'
# select_model = 'resnet152'
# select_model = 'vgg11'
# select_model = 'vgg16'
# select_model = 'vgg19'
# select_model = 'efficientnetb0'
# select_model = 'efficientnetb5'
select_model = 'mobilenetv2'
###[INFO] set num of iterations to run (train.py) on (run_train.sh) file /// this automatically saves training info for every 20 epochs.
bash run_train.sh

recommended new environment to infer models only on cpu

for testing

conda create -n test2 python=3.10 -y
conda activate test2
pip install tensorflow ultralytics
python demo.py

realtime BEV plot

set
## BEV_plot = True
## TracK = True

Contributing

Contributions are welcome! If you find any issues or have suggestions for improvements, please open an issue or submit a pull request.

Acknowledgements

Expand