PJYoloVision: Object Detection and Instance Segmentation AI project

Note

Next version will include models in jax/flax

PJYoloVision is an AI project that includes state-of-the-art object detection and instance segmentation models. The models are part of the YOLO (You Only Look Once) family, which is a group of models that use deep learning to perform real-time object detection, classification, and localization.

Object Detection Models

The object detection models in PJYoloVision can detect and locate objects in real-time images and video streams. These models can be trained to detect specific types of objects, such as cars, people, animals, or any other object of interest. The YOLO models are known for being fast and accurate, making them a good choice for real-time applications.

Instance Segmentation Models

The instance segmentation models in PJYoloVision can detect and locate individual objects within an image or video frame. Unlike object detection models, instance segmentation models also provide a mask for each object, which can be used to separate them from the background. This makes them ideal for applications like medical imaging, where precise segmentation is essential.

YOLO Family

The YOLO family includes several models, each with their own strengths and weaknesses. YOLOv1 was the first model in the family, and it introduced the concept of object detection using a single neural network. YOLOv2 improved upon the original model by incorporating a more powerful backbone network and making it easier to train. YOLOv3 further improved on the model by incorporating new techniques such as feature pyramid networks and improved training strategies.

Customizable Models

With PJYoloVision, users can create their own custom models by modifying the configuration files and training them with their own datasets. This allows users to build models tailored to their specific needs and applications.

Special Models

In addition to the standard object detection and instance segmentation models, PJYoloVision also includes models for other types of computer vision tasks. For example, there is a model for determining the age and gender of people in images, as well as a model for detecting emotions.

Overall

PJYoloVision is a powerful AI project that provides a variety of high-quality models for object detection and instance segmentation tasks. The models are part of the YOLO family, which is known for being fast and accurate, and they can be easily customized to meet the needs of specific applications.

Note

A Part of the idea and source code is from YOLO v8 ultralytics and PJYoloVision is an edited version of YOLO v8 (a weak version for education and self learning) and CLI is not perfect (it's not working) if you want to use this project you are free this project is GPL-3.0 license licenced, but it's recommended to use ultralytics for more stability

Usage Note

PJYoloVision currently share same api as ultralytics YOLOv8 so you can predict use and train model via same api as ultralytics

Example

## Use normal detection model 
from PJYoloVision import YOLO

# to build model from yaml file just pass yaml file (same as ultralytics)
model = YOLO('MODEL.pt')
image = 'your image (path , url , numpy array ...)'
predicts = model.predict(image, save=True)

Segmentation

## Use Segmentation model 
from PJYoloVision import YOLO, prepare_segmentation

# to build model from yaml file just pass yaml file (same as ultralytics)
model = prepare_segmentation(YOLO('MODEL-Segmentation.pt'))
image = 'your image (path , url , numpy array ...)'
predicts = model.predict(image, save=True)

If you want to use segmentation models you have to set the task manually for v 0.0.1 models

from PJYoloVision import YOLO, prepare_segmentation

model = prepare_segmentation(YOLO('PJPJYoloVision-Segmentaion-M.pt'))
## Ready to Use 

for more setting and access to the model you can use set_new_setting

from PJYoloVision import YOLO, set_new_setting

# to build model from yaml file just pass yaml file (same as ultralytics)
model_args = dict(amp=False)
model = set_new_setting(YOLO('MODEL.pt'), **model_args)
image = 'your image (path , url , numpy array ...)'
predicts = model.predict(image, save=True)