search

RuojinCai / doppelgangers

Doppelgangers: Learning to Disambiguate Images of Similar Structures
https://doppelgangers-3d.github.io/
MIT License
173 stars 23 forks source link

Dense reconstruction #11

Open wdstar1 opened 10 months ago

wdstar1 commented 10 months ago

Hello, first of all congratulations on your excellent work, I used doppelgangers for sparse reconstruction of 35 photos and then used colmap for dense reconstruction, but the number of images that could be densely reconstructed at this time was only 3. I would appreciate it if you could help me with this problem. My threshold is set to 0, and in my understanding it should not filter out any image pairs.

RuojinCai commented 1 day ago

Hello, I hope the question has been resolved. If not, have you checked how many pairs were deleted when you set the threshold to 0? It should print a message like 'Total', cursor.rowcount, 'Records deleted successfully' (https://github.com/RuojinCai/doppelgangers/blob/main/doppelgangers/utils/process_database.py#L50). For a sanity check, did you check, without running the doppelganger algorithm, if all 35 images can be registered in dense reconstruction?