X-ray Diffraction Pairs for Noise Filtering Algorithm Training

Jens Oppliger's dataset from the University of Zurich provides paired low- and high-counting X-ray diffraction frames for training noise filtering algorithms. The data was recorded on a La1.88Sr0.12CuO4 single crystal at the PETRA III storage ring at DESY in Hamburg, using a Dectris Pilatus 100K CdTe detector with 100 keV photons and the sample cooled to ~30 K. It contains sequences of 41 consecutive frames per scan, representing signals like 2D charge density wave order, Bragg peaks, powder lines, spurions, and dead pixels.

Use Cases

• Training neural networks for experimental noise reduction based on paired low/high-count image sequences.
• Benchmarking denoising algorithms on real X-ray diffraction data containing mixed signals like Bragg peaks and charge density waves.
• Studying the characteristics of noise in synchrotron-based diffraction measurements from the specified detector and beamline setup.

Strengths

• Contains more than 200 groups of paired low- and high-counting statistics, providing a substantial basis for supervised learning.
• Includes detailed experimental metadata such as sampling times, monitor values, and specific detector and beamline conditions.
• Data is structured into training, validation, and test sets, with the test set including pre-processed denoised frames from a neural network.

Limitations

• Conflicting exposure times are reported across sources (1s/20s vs. 2s/21s for low/high counts), creating uncertainty about precise acquisition parameters.
• Key metadata such as total row count, file size, and last update date are missing from all platform entries.
• The dataset is highly specific to a single material and experimental setup, limiting generalizability to other systems.

Provenance

Source

University of Zurich; data recorded at beamline P21.1, PETRA III storage ring, DESY, Hamburg, Germany.

Collection Method

Experimental measurement using a Dectris Pilatus 100K CdTe detector on a La1.88Sr0.12CuO4 single crystal.

Geography

Hamburg, Germany (experimental site).