Pulse Shape Discrimination Algorithms: A Survey and Benchmark – Datasets, Pre-trained Models, and Results
收藏资源简介：
This Zenodo bundle accompanies the open-source "Pulse Shape Discrimination (PSD)" toolbox hosted on GitHub. It contains all data, pre-computed results, and parameter files required to reproduce the figures, tables, and benchmarks reported in our survey paper: "Pulse Shape Discrimination Algorithms: A Survey and Benchmark" by Haoran Liu et al. (2025). If you use any part of this work, please cite both the original paper and this Zenodo dataset. **Repository Contents:** The repository is organized into the following ZIP archives: * **Datasets.zip**: Contains the raw pulse waveform data. * **AmBe**: Data from the Americium-Beryllium source, including training, validation, test, and reference signal sets. * **PuBe**: Data from the Plutonium-Beryllium source at 2, 3, 4, and 5 MeV, each with corresponding training, validation, and test sets. This data includes ground-truth labels from time-of-flight (TOF) measurements. * **Experimental Results AmBe.zip**: Pre-computed models and performance metrics for the AmBe dataset. * **Experimental Results PuBe [2-5] MeV.zip**: Four archives containing pre-computed results and models for the PuBe dataset, one for each energy level (2, 3, 4, and 5 MeV). Results are categorized by method: * **Deep Learning**: Trained models (`.pth`), validation results, and summaries. * **Machine Learning**: Trained models (`.joblib`, `.pkl`), validation results, and summaries. * **Statistical**: Validation results (FOM, ROC curves) and summaries. * **PSD Parameters for PuBe Dataset.zip**: Python script files with algorithm hyperparameters tuned specifically for the PuBe dataset. **Usage:** This dataset is designed for use with the open-source PSD toolbox on GitHub. To reproduce our results, clone the repository, unzip the archives from this bundle, and follow the instructions in the included `README.txt` file. **License:** All files are released under the MIT License. **Contact:** For questions or issues, please contact liuhaoran@cdut.edu.cn or open an issue on the project's GitHub repository.
