Laser Resonator Beam Alignment: Dataset (Subset)

This dataset supports the study “A Two-Stage Learning Framework with a Beam Dataset for Laser Resonator Alignment”. It includes 500 resized PNG beam images (256×256), a labels.json file recording alignment parameters, a sampled_pairs_500K.json listing 500 000 image pairs, and a README describing structure and usage. Full-resolution raw data will be released upon publication.
In addition, a demo video of mp4 type is included, called "Demo Video Dataset Evaluation and Real world Resonator Test".
This video consists of two parts.
The first part shows performance evaluation on the proposed beam image dataset, demonstrating alignment inference accuracy and parameter prediction.
The second part presents real-world testing results using a lab-based, industry-inspired laser resonator setup, showcasing automatic beam adjustment driven by our model’s inference.