Object Detective: A Desktop App

Current version: Alpha 0.95
Note: Application is currently in Turkish language. There'll be English version in final. For Turkish documentation click here!

About

Goal: Detect track and count the object according to its horizontal direction.
Object Detective is an open-source, cross-platform desktop application written in Python. It uses MySQL server to store its data and it requires a login system.
You can use this code for your own purposes however if you need technical support, please consult us. You can reach out via email or LinkedIn for the right price and customized version for your purpose of our app.
Currently our app supports for one-class object detection (which is cows right now) however anyone who is eligible enough, can train their own TensorFlow model and use this algorithm with it.

Team

How To Run?

Suggested Python version is 3.6 or higher.

Create a MySQL server from file named detectionDB_v1.sql and configure your sign_in/db_connection file according to your server.
Install requirements.
Make sure cmake and dlib is installed.
cd ./app (Go to the file named app)
Run program.py!

For Windows: python program.py
For Linux: sudo python3 program.py (Tested on Ubuntu 20.04) -- Not tested on MacOS

Special thanks to:

pyimagesearch.com for sharing the tracking algorithm for free thus we customized for TensorFlow models.
Wanderson Magalhaes for amazing GUI design and ideas.

Photos

Main window (Not signed in - Watch out profile icon at left bottom-):
Active main window (signed in):
Logs:
Signed in user window (PC ID attribute is a placeholder right now. Still working on it.) :
Sign in window:

Final notes: Soon my colleagues will be sharing the cow dataset that we pictured and organized for training the model.