sadi-ai's Introduction

SADI A.I.

A.I. powered security systems could lead us to a safer world. Real-Time face detection, threat detection, face recognition, and individual detection are some main functionalities of this system.

Functionalities

Pretrained:

You can find them here

Person detection:

This model named

person_detection.pt

has been made for person tracking.

Threat detection:

This model named

threat_detection.pt

has been made for threat tracking.

Person and threat detection:

This model named

person_and_threat_detection.pt

has been made for person and threat tracking.

To train

face recognition:

This model has been made for you to give it a dataset, and then automates the training process to get the model.

In order to train it, follow these steps:

Give a dataset: For face recognition, we want labels/classes where the photos of individuals are.

The directory where you should put the classes images is: ./training/face_models/YOLO_data_utils/data/

    ... data/
            ---individual_1:
                photos
            ---individual_2:
                photos
            ---individual_3:
                photos
            ---individual_4:
            ...
            ...
            ...
            ...
            ...

The process is detailed on the README.md inside data folder.

Run main.py on face_models.
Wait until the model finishes the training phase.
Move the best.pt model from the training session you accomplished and send it to the /models/ folder.
Now you can access to the model and save it on /models/ directory, there is a preview executing app.py

Recommend Projects