yjhuasheng / cnn_palmar_veins

Web application - palm veins recognition

Project Palm veins recognition

By Kenan, Fayçal, and Lorenzo - Engineering students of the school ESME Sudria

Table of Contents

1. About The Project
2. Getting Started
   • Installation
3. To Run Docker
4. Contact

About The Project

GOAL : Our goal is to train a model to recognize the users of the palm veins. "Palm veins" are the veins that are located on the palm of the hand and is unique to everyone like a fingerprint.

Here are some of the model we have trained without transfer learning:

Model	Accuracy	Loss
1 Layer	99.11%	0.09
ResNet 50	99.22%	0.15
ResNet 150	96.11%	1.87
EfficientNetB0	99.22%	0.19
Xception	99.89%	0.01
InceptionResNet	99.67%	0.01

Getting Started - Web App and Docker

Installation

Docker

1. Install docker, and preferably also install docker-desktop.

2. Now you need to install docker compose, to do so run the following command to dl:

   sudo curl -L "https://github.com/docker/compose/releases/download/1.29.2/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose

3. Now we need to apply executable permissions to the binary to allow us to run compose

   sudo chmod +x /usr/local/bin/docker-compose

To Run Docker

Type in the terminal (where the docker-compose.yml is), to run the project :

docker-compose up -d

This will launch 1 containers : Python-Flask which is located at the directory "app". (app.py file is the main file of the web application)

When every containers is running like this:

Go to : http://localhost:5000/

Now you can see the web app running. The application is running on port 5000 on mode debug and the web app is accessible.

You can go to the "palm veins identification" page.

Then you can upload your palm vein image.

When you press the button "Prediction" you will see the result of the model.

To shutdown docker

When you have finished using the application :

To shutdown all the containers

docker-compose down

To build all the images of your project docker

To rebuild all your images Docker

docker-compose build

Contact

Student:

• Kenan
• Fayçal
• Lorenzo

Tuteur : Larbi BOUBCHIR

Project link: Palm veins recognition