In the EU-funded LEAPS-INNOV project, large photon science facilities are collaborating to develop innovative technologies relevant to their experiments. One workpackage focuses on data reduction techniques for experiments such as X-ray diffraction and tomography. This work covers three main areas:

• Developing and evaluating algorithms for data reduction. This has included, for instance, evaluating image denoising algorithms to both improve image quality and make images more compressible, and investigating machine-learning-based optimization of compression parameters.
• Disseminating knowledge on data reduction by running online seminars on data reduction in different areas of X-ray science, and organizing events such as a summit on compression in the HDF5 file format.
• Improving software libraries and tools for data reduction. Recently, in consultation with the company ironArray SLU, we integrated new features into the HDF5 file format such as JPEG2000 compression, to make it possible to more conveniently store large compressed datasets along with metadata.