sdsubhajitdas / Multi-Label-Image-Classification-Demo

B.Tech Final year project demo

Multi Label Image Classification Demo

A simple and functional demo website for my B.Tech final year project demo.

Getting started.

You need to have python3 and pip3 in your system.

Installing

• First clone the project

  git clone https://github.com/sdsubhajitdas/Multi-Label-Image-Classification-Demo.git
  

• (Optional) If you like you can use virtualenv to create a private environment and then activate it.
• Installing the required packages.

  pip install -r requirements.txt
  

Usage

• Running the project is very simple. It is a simple flask application which is serving a tensorflow model already pretrained & included in the repository.
  Run this command to start the local server.

  python run.py
  

  This should give a output like this.
  

• Copy the url shown in the terminal and open in your browser. You will be greeted with a page like this.

Outputs

Our demo can identify 20 different classes of objects. Below are the following classes.

Aeroplane, Bicycle, Bird, Boat, Bottle, Bus, Car, Cat, Chair, Cow, Dining Table,
Dog, Horse, Motorbike, Person, Potted Plant, Sheep, Sofa, Train, Tvmonitor

If you are interested about all the training, model generation work, what methods and architeture we have used then you will have to wait. Will publish the other repository soon here (No link yet).

Some of the results are below. Click to see the outputs.

Output 1

Output 2

Output 3

Output 4

Output 5

Output 6

Output 7

Contributors

• Adrika Nandi
• Angana Bose
• Debolina Lahiri
• Ishani Banerjee
• Subhajit Das
• Urmila Saha

License

This project is licensed under the MIT License - see the LICENSE file for details