COMP 472 Project

"SmartClass A.I.ssistant", CNN implementation that can analyze facial expressions and categorize them into different states/classes.

Group Members

• 40231530, Bryan Carlo Miguel
• 40212780, Yasser Ameer
• 40213100, William Nazarian

Table of Contents

1. Setting up the Project
2. Executing Data Visualization Scripts
3. Project Contents

Setting up the Project

The project comes with a requirements.txt for installing dependencies. The simplest way to get the program running is to create a venv, and then install the dependencies as seen below:

$ py -m venv comp472
$ comp472/Scripts/activate
$ pip install -r requirements.txt
    ...

Executing Data Visualization Scripts

The project splits the visualization functionality into multiple scripts that can take input parameters.

Class Distribution

The module scripts.plot_bars contains the functionality for visualizing the number of images in each class. Executing the following command executes the script.

$ py -m scripts.plot_bars
    ...

Pixel Intensity Distribution

The command py -m scripts.plot_histograms <CLASS NAME> visualizes the distribution for the specified class. For example, the following command visualizes the intensities for the class happy.

$ py -m scripts.plot_histograms happy 
    ...

Sample Images

The command py -m scripts.image_sampler <CLASS NAME> samples 15 images from the specified class, plots each of the images' pixel intensities, then displays them all side-by-side in a 5x3 grid. For example, the following command does this sampling and visualization for the class happy.

$ py -m scripts.image_sampler happy 
    ...

Train and Evaluate CNN Model

To train and evaluate the model first run the following command in the root of our project

python main.py

Then you will be asked to choose one of our different types of models to train the facial expression images

1. Main Model
2. Variant 1
3. Variant 2

After picking you will be asked to name the CNN to whatever you would like. Finally, you get the see the training process and how the model improves after each epoch reducing the value of its loss function and increasing the accuracy

Project Contents

• The part1 directory contains the dataset, as well as a .csv file containing paths to the raw images, as well as their classification.
• The scripts directory contains the main scripts used for data cleaning and visualization.
• The utils directory contains any supporting python modules that contain common functionality.

project directory
├── part1 
│   ├── structured_data 
│   ├── Combined_Labels_DataFrame.csv 
│   │    ... 
│
├── scripts 
│   │    ... 
│
└── utils 
    │    ...