fmap 🧠🚀

About Us 👥

We are a team of four passionate machine learning students from Politecnico di Milano, Italy. Our team, famp, participated in the challenges proposed by the Artificial Neural Networks and Deep Learning course.

Team Members:

• Filippo Galli
• Alessandro Howe
• Matteo Callini
• Paolo Bellezza

Project Highlights 🏆

🔬 Challenge 1: Blood Cell Classification

Dive into Medical Image Analysis

• Goal: Classify blood cell images into 8 distinct types
• Key Achievements:
  • Developed advanced deep learning models
  • Achieved 98.75% accuracy on local tests
  • Implemented sophisticated data augmentation techniques

🌍 Challenge 2: Mars Terrain Segmentation

Explore Extraterrestrial Landscapes

• Goal: Segment Martian surface images into 5 terrain classes
• Key Achievements:
  • Created innovative neural network architectures
  • Developed U-Net and U-Net++ models
  • Achieved 64.05% validation Mean IoU

Repository Structure 📂

famp/
│
├── challenge_1/        # Blood Cell Classification
│   ├── fmap.ipynb      # Final model notebook
|   ├── fmap_report.pdf # Final report
│   ├── README.md       # Detailed project documentation
│   └── aux_notebooks/  # Additional materials
│
├── challenge_2/        # Mars Terrain Segmentation
│   ├── fmap.ipynb      # Final model notebook
|   ├── fmap_report.pdf # Final report
│   ├── README.md       # Detailed project documentation
│   └── aux_notebooks/  # Additional materials
│
└── README.md          # Main repository overview

Technologies Explored 💻

Getting Started 🚀

Prerequisites

• Python 3.x
• TensorFlow 2.x
• Keras
• Jupyter Notebook

Installation

1. Clone the repository

   git clone https://github.com/your-username/FMAP-ML-Challenges.git

2. Navigate to the challenge folder

   cd FMAP-ML-Challenges/challenge_1  # or challenge_2

Warning

We didn't make a requirements.txt file since all the test was done in Colab or Kaggle.

Detailed Exploration 🔍

Each challenge folder contains:

• Comprehensive Jupyter Notebook with model implementation
• Detailed README explaining the project
• Project reports and additional documentation

Recommended Reading Order

1. Start with the main README
2. Explore individual challenge READMEs
3. Deep dive into Jupyter Notebooks
4. Review project reports

Learning Insights 💡

These projects showcase:

• Advanced transfer learning techniques
• Sophisticated data augmentation
• Model optimization strategies
• Handling real-world machine learning challenges

Contributions Welcome! 🤝

Interested in contributing or learning more?

• Open issues for questions
• Submit pull requests
• Provide feedback

Connect With Us

• Filippo Galli: LinkedIn
• Alessandro Howe: LinkedIn
• Matteo Callini: LinkedIn
• Paolo Bellezza: LinkedIn

Acknowledgments

Special thanks to Politecnico di Milano and the Artificial Neural Networks and Deep Learning course instructors.