Attire Classifier is a Cultural Clothes Classifier aimed at creating a CNN Model that can classify clothes from different regions of Pakistan.
Following Dataset was assembled and used: https://www.kaggle.com/datasets/khizertariq/pakistani-cultural-clothes
For classification a CNN Model was used that predicts outputs for four classes.
The converter.ipynb can be used to convert files to jpg format while the datacleaningscript.ipynb can be used to clean the directories while keeping images of only the allowed formats.
The Code also uses Gradient-weighted Class Activation Mapping (Grad-CAM) to highlight the regions influencing the decisions of the CNN model thus taking a step towards Explainable AI.