Comet Centroiding Benchmark: Astrometric & Morphological Validation Suite

Official benchmarking and validation suite designed to rigorously evaluate and compare astrometric centroiding algorithms against complex cometary morphologies, including jets, diffuse comae, background gradients, and low-SNR regimes.

The results and methodology validated through this benchmark inform the core algorithms implemented in Kometra (developed within the research activities at PoliTO Astronomy).

---

1. Overview

Traditional stellar centroiding algorithms, such as DAOPhot, standard center-of-mass methods, or basic Gaussian fits, can fail or introduce severe systematic biases when applied to cometary targets due to their non-Gaussian profiles, active jets, diffuse comae, and dense stellar fields.

This repository provides:

1. The Generation Pipeline Python scripts to synthesize realistic FITS images with known ground-truth coordinates (true_x, true_y).

2. The Benchmark Engine Automated testing frameworks to evaluate centroiding accuracy, robustness, window-size sensitivity, and offset drift.

3. The Official Dataset A pre-generated, standardized validation suite openly available through Zenodo.

---

2. Repository Structure

comet-centroiding-benchmark/
├── src/                    # Core source code for simulation, benchmark, and analysis
├── data/                   # Local working directory for datasets and ground truth
├── results/                # Output directory for benchmark data and analysis reports
├── 01_generate_dataset.py  # Generate synthetic FITS images and CSV annotations
├── 02_run_benchmark.py     # Execute the centroiding benchmark tests
├── 03_analyze_results.py   # Generate statistical reports and validation plots
├── run.py                  # Master control script for the pipeline
├── pyproject.toml          # Package configuration and dependencies
└── LICENSE                 # GNU General Public License v3.0

---

3. Installation

Clone the repository and install the package in editable mode together with its dependencies:

git clone https://github.com/Genofabio/comet-centroiding-benchmark.git
cd comet-centroiding-benchmark
pip install -e .

---

4. Usage: Running with the Official Zenodo Dataset

For reproducible benchmarking, it is recommended to use the official standardized test suite published on Zenodo.

Dataset Reference: Genovese, F. (2026). Astrometric and Morphological Validation Dataset for Kometra (Synthetic Comet Benchmark Suite). Zenodo. DOI: 10.5281/zenodo.22014531

Setup Steps

1. Download and extract the dataset archive from Zenodo.

2. Place the contents of main_benchmark/ inside:

data/synthetic_test/

3. Copy synthetic_ground_truth.csv into:

data/

4. Execute the evaluation and analysis steps, skipping dataset generation:

python run.py benchmark
python run.py analyze

---

5. Usage: Generating the Dataset from Scratch

If you prefer to generate the synthetic images locally using the built-in physical models, run the complete pipeline through the master script.

The complete workflow consists of:

Generation → Benchmark → Analysis

python run.py all

The individual pipeline components are also available through the dedicated scripts:

python 01_generate_dataset.py
python 02_run_benchmark.py
python 03_analyze_results.py

---

6. Citation

If you use Kometra, its benchmark suite, or the associated dataset in your research, please cite both the software repository and the official dataset.

Dataset Citation

@dataset{genovese_2026_dataset,
  author       = {Genovese, Fabio},
  title        = {{Synthetic Cometary Dataset for Centroiding Algorithm Selection (Kometra Benchmark Suite)}},
  month        = aug,
  year         = 2026,
  publisher    = {Zenodo},
  version      = {1.0.0},
  doi          = {10.5281/zenodo.22014532},
  url          = {[https://doi.org/10.5281/zenodo.22014531](https://doi.org/10.5281/zenodo.22014531)}
}

---

7. License

This project is open-source software licensed under the GNU General Public License v3.0 (GPL-3.0).

See the LICENSE file for the complete license terms.

---

8. Reproducibility

The official Zenodo dataset provides a fixed and versioned benchmark reference for reproducible evaluation.

For published research, it is recommended to report:

• the Kometra Benchmark version;
• the Zenodo dataset version;
• the benchmark configuration;
• the centroiding methods evaluated;
• the ROI/window sizes used;
• the resulting astrometric error metrics.

This ensures that benchmark results can be independently reproduced and compared across implementations.

---

9. Project Links

• Source repository (Benchmark): https://github.com/Genofabio/comet-centroiding-benchmark
• Kometra (Core Software): https://github.com/Genofabio/kometra
• PoliTO Astronomy Fork: https://github.com/PoliTO-Astronomy/Kometra
• Official dataset (Zenodo): https://doi.org/10.5281/zenodo.22014531
• License: https://www.gnu.org/licenses/gpl-3.0