dtourolle/DarkRoom
1
0
Fork 0
Code Issues 39 Pull Requests Actions Packages Projects Releases 20 Wiki Activity

Dump RAW mosaics for training the learned denoise

Browse Source 
denoise.md §4.4 requires the training repo to read photosites through
dr-decode, not LibRaw, so black and white levels, the active area and the
CFA phase match what the app will feed the network. mosaic_dump reads
`input<TAB>prefix` lines and writes the whole readout as .npy plus a JSON
of what decode and metadata report. The masked border is kept: its
optically black photosites are a free dark frame for the noise profile.
This commit is contained in:
dtourolle
2026-10-03 10:25:27 -04:00
parent 2fad846cd1
commit 1f266a4478
2 changed files with 100 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/dev/traceability.md
+1 -1
View File
@@ -9,7 +9,7 @@ Denominators are parsed from [`requirements.md`](requirements.md) at run time, n
| Metric | Value |
|---|---|
| Source files scanned | 486 |
| Source files scanned | 487 |
| TRACES tags found | 2033 |
| Requirements defined | 193 |
| Requirements deferred (post-v1) | 24 |