Hi,
First of all, thank you for the great work on this dataset!
While analyzing a subset of the thermal images, I noticed an unusual distribution in the pixel values (see histogram below). The pixel values on the raw published tiff images are in the 0–255 range, but the histogram shows distinct clustering/accumulation in certain bins/values with a repeating pattern, as if some values are disproportionately represented. This makes me suspect that some kind of quantization, rounding, or scaling might be affecting the raw data.
Could you clarify whether the thermal images are post-processed or rescaled in any way before being included in the dataset? Understanding this would really help in interpreting the thermal data correctly.
Thanks in advance for your insights!
Hi,
First of all, thank you for the great work on this dataset!
While analyzing a subset of the thermal images, I noticed an unusual distribution in the pixel values (see histogram below). The pixel values on the raw published tiff images are in the 0–255 range, but the histogram shows distinct clustering/accumulation in certain bins/values with a repeating pattern, as if some values are disproportionately represented. This makes me suspect that some kind of quantization, rounding, or scaling might be affecting the raw data.
Could you clarify whether the thermal images are post-processed or rescaled in any way before being included in the dataset? Understanding this would really help in interpreting the thermal data correctly.
Thanks in advance for your insights!