search

DY112 / LSMI-dataset

[ICCV'21] Large Scale Multi-Illuminant (LSMI) Dataset for Developing White Balance Algorithm under Mixed Illumination
https://www.dykim.me/projects/lsmi
58 stars 3 forks source link

The first step, tiff image generation problem #25

Open xiaosanxiao13 opened 2 weeks ago

xiaosanxiao13 commented 2 weeks ago

When I used 0_cvt2tiff.py, I found that all the generated tiff images were almost pure black. After experiments, I found that cmd = [DCRAW_PATH, "-h", "-D", "-4", "-T ", full_path], what I want to know is, is there any problem with the pure black tiff image generated in this way? Is this the case for everyone?

DY112 commented 1 week ago

Hello, thank you for your interest in my work The converted TIFF images are 16bit uncompressed and contain raw data. The maximum value of RAW data depends on the camera, but Galaxy is 1024 (10bit) and Sony or Nikon is 14bit. Therefore, the thumbnails of TIFF images may look black. You can rescale them with MAX value or gamma correction to visualise them well.

If you open the black tiff image with the cv2.imread function using the cv2.IMREAD_UNCHANGED tag, you will see that the value is not zero.