Link for the saved models for Part III:

• 10-fold: https://drive.google.com/drive/folders/1V_UoOJDh0mqlWAUu6MjPlJDZ185xhV3T?usp=share_link
• Bias: https://github.com/JRB958/COMP472-PROJECT/blob/main/nonbias_model.zip

COMP 472 Artificial Intelligence

Team Name: OB_03

Team Members:

• Gevorg Alaverdyan (#40202177) - Data Specialist
• Jay Patel (#40203705) - Training Specialist
• Joud Babik (#40031039) - Evaluation Specialist

Content Description

The /scripts folder contains scripts for data cleaning and visualization:

• scripts/image_names_getter.py: Retrieves file names and inputs them into an Excel file.
• scripts/image_resize.py: Removes any image smaller than 10KB and resizes images to 224x224 pixels.
• scripts/aggregated_histograms.py: Creates aggregated histograms for all images in a directory and its subdirectories.
• scripts/Random_Histograms_Per_Folder.py: Selects 15 random photos from a directory, generates histograms for them, and saves the histograms along with the images in a specified directory.

Steps to Execute Code

a) Data Cleaning Scripts

image_names_getter.py:

1. Copy and paste this script into your IDE of choice.
2. Update the variables:
   • root_directory = "/Users/username/..." (Set to your root directory)
   • excel_path = "/Users/username/.../image_references_focused.xlsx" (Set to your desired output Excel file name)
3. Run the script.

image_resize.py:

1. Copy and paste this script into your IDE of choice.
2. Update the variables:
   • file_path = "C:\\Users\\gevor\\Downloads\\colored2" (Set to your input directory. Ensure this directory exists before running the script.)
   • outfile = "C:\\Users\\gevor\\Downloads\\fixed\\" (Set to your desired output directory. Ensure this directory exists before running the script.)
3. Run the script.

b) Data Visualization Scripts

aggregated_histograms.py:

1. Copy and paste this script into your IDE of choice.
2. Run the script.
3. When prompted, paste the absolute path of the directory you want to aggregate pixel intensity for.

Random_Histograms_Per_Folder.py:

1. Copy and paste this script into your IDE of choice.
2. Update the paths:
   • Set the relative path to your source directory on line 207.
   • Set the relative path to your destination directory on line 209.
3. Run the script.

b) Training and Evalution script

CNN_model.py:

1. Copy and paste this script into your IDE of choice.
2. set the path of the data set (line 22)
3. set the path of the best model to be saved (line 23)
4. set the path of the csv that will contain the metrix (line 373)
5. Run the script.