Image-based food recognition with U-Nets

The repository contains the project realized for the Deep Learning course of the Master's degree in Artificial Intelligence, at Alma Mater Studiorum, University of Bologna.

The projects is realized by:

• Alex Costanzino (arcanoXIII), [email protected];
• Marco Costante (Markostante), [email protected].

Contents

• 01_dimension_check.py and 02_reordering_dataset.py are just auxiliary tools to pre-process the dataset;
• u_net_for_image-based_food_segmentation.ipynb is the main notebook;
• deep_u_net_for_image-based_food_segmentation.py is the alternative architecture;
• demo_evaluation.ipynb is a notebook for a quick evaluation of the various models;
• report.pdf is the report file.

Main libraries

• TensorFlow with Keras backend;
• Numpy 1.20.0;
• Pandas 1.2.4;
• OpenCV 4.5.2;
• PIL 8.2.0;
• pycocotools;
• tqdm 4.61.0;
• matplotlib 3.4.2;
• segmentation_models.

Further details can be found in the report.

License

This project is licensed under the MIT License - see the LICENSE file for further details.