Multi-label Image Segmentation with Fashion Images

The purpose of the project was to segment clothing articles from images by masking each pixel into its respective categories.

In the Exploratory Data Analysis notebook, I analyze images in the dataset and identify issues that may affect the training of the model.

In the model training notebook, I choose the appropriate encoding method, model architecture, and configuration. At the end of the notebook, I analyze the result and provide recommendations to further improve the model.

The notebook is divided into 2 parts:

• Exploratory Data Analysis on the images, available here

• Model Training and and findings, available here