This repository implements machine learning models for Non-Line-of-Sight (NLOS) object detection — focusing on both classification (identifying the object) and localization (pinpointing its position) using indirect signal data.
- Develop models capable of detecting and localizing objects hidden from direct view using NLOS data.
- Compare multiple modeling strategies for classification and localization.
- Provide reusable code, experiment notebooks, and a pre-trained checkpoint for quick inference.
nlos-object-detection-localization/
├── data/ # Includes input data and preprocessing scripts
├── model.ipynb # Core Jupyter notebook with modeling implementation
├── simple_data_test.ipynb # Notebook for preliminary data exploration & testing
├── reg_cls_checkpoint.pt # Pretrained checkpoint (classification/localization)
└── README.md # Project documentation