Electron-Photon Classifier

This project implements a deep learning model for classifying electrons and photons using a custom ResNet15 architecture. The model analyzes calorimeter data from particle physics experiments to distinguish between electron and photon signatures.

Project Overview

The classifier uses a ResNet15 architecture with the following key features:

• Custom residual blocks with batch normalization
• Two input channels (ECAL and HCAL calorimeter data)
• Achieves 73.39% accuracy on test set
• ROC AUC score of 0.80

Project Structure

.
├── models/
│   ├── __init__.py
│   ├── block.py         # Implementation of ResNet building blocks
│   └── models.py        # ResNet15_v1 and ResNet15_v2 architectures
├── train/
│   ├── __init__.py
│   ├── evaluate.py      # Model evaluation utilities
│   └── train.py        # Training loop and scheduler implementations
├── utils/
│   ├── __init__.py
│   ├── data_loader.py   # Dataset loading and preprocessing
│   ├── pre_processing.py # Data normalization
│   └── visualization.py  # Plotting utilities for results
├── main.py             # Training script
├── inference.py        # Inference script
└── main.ipynb         # Interactive notebook for training/testing

Prerequisites

• Python 3.7+
• PyTorch
• h5py
• numpy
• scikit-learn
• matplotlib
• torchsummary

Installation

1. Clone the repository:

git clone https://github.com/JDhruvR/ElectronPhotonClassifier.git
cd ElectronPhotonClassifier

2. Download the dataset:

mkdir data
cd data
curl -o SinglePhoton249k.hdf5 https://cernbox.cern.ch/remote.php/dav/public-files/AtBT8y4MiQYFcgc/SinglePhotonPt50_IMGCROPS_n249k_RHv1.hdf5
curl -o SingleElectron249k.hdf5 https://cernbox.cern.ch/remote.php/dav/public-files/FbXw3V4XNyYB3oA/SingleElectronPt50_IMGCROPS_n249k_RHv1.hdf5
cd ..

Usage

Training

python main.py

Inference

python inference.py

For interactive development, you can use main.ipynb which contains both training and testing code. This notebook can be run directly in Google Colab.

Model Architecture

The ResNet15_v2 architecture includes:

• Initial convolution layer with 32 filters
• 5 residual blocks with increasing channel dimensions (32→64→128→256→512→1024)
• Batch normalization and ReLU activation throughout
• Final fully connected layer for binary classification

Training

The model is trained with:

• Adam optimizer
• Cross-entropy loss
• Learning rate: 1e-3
• Weight decay: 1e-4
• Plateau learning rate scheduler
• 30 epochs

Results

The model achieves:

• Accuracy: 73.39%
• ROC AUC Score: 0.80

Our implementation replicates the predictions from the E2E CMS paper.