Missing Metadata in Benchmark Datasets

[bkhanijo]

I am working with the sampled datasets provided in the benchmark and noticed some critical metadata is missing or inconsistent. I would like to request the following details to ensure reproducibility and accurate usage of the data:

1. m-Neon-Tree Segmentation Dataset :

• Spectral Wavelengths : The original Neon Tree dataset contained larger number of bands. Could you please provide specific wavelengths for subset of 369 bands (in the Neon band in the dataset).
• Scaling Algorithm Used : Also, would request you to share the scaling algorithm used across the benchmark to resize information across bands in case their spatial resolution doesn't match. This is because the spatial resolution of the data shared across bands in Neon-Tree dataset is different (1m vs 0.1m).

2. Geolocation information :

• Transform Matrix: The value in transform field is either missing or same for samples of a dataset. Each sample should ideally have a unique transform to reflect its specific spatial location.

• CRS/EPSG Code : The value of Coordinate Reference System (CRS) or EPSG code is also missing in certain datasets.

Providing this information would be greatly appreciated to facilitate proper georeferencing and analysis. Thanks in advance!