Initial tests - Large Scenes - Tanks and Temples, Bureau (HDLM), building

Lay3rss / gaussian-opacity-fields

0 stars 0 forks source link

Initial tests - Large Scenes - Tanks and Temples, Bureau (HDLM), building #1

Open rebecca-lay3rs opened 1 month ago

rebecca-lay3rs commented 1 month ago

observation

to run training, activate the conda environment: conda activate gof
from root, enter the gof folder: cd gaussian-opacity-fields/

Barn (Tanks and Temples - TnT)

training

-r 2 for using downsampled images with factor 2
--use_decoupled_appearance to enable decoupled appearance modeling if your images has changing lighting conditions

python train.py -s TNT_GOF/TrainingSet/Barn -m exp_TNT/Barn -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m exp_TNT/Barn --iteration 30000

rebecca-lay3rs commented 1 month ago

Bureau (761 imgs, complete room)

training

python train.py -s data/Bureau_GS_data -m outputs/Bureau -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m outputs/Bureau --iteration 30000

rebecca-lay3rs commented 1 month ago

Courthouse (Tanks and Temples - TnT)

training

python train.py -s TNT_GOF/TrainingSet/Courthouse -m exp_TNT/Courthouse -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m exp_TNT/Courthouse --iteration 30000

rebecca-lay3rs commented 1 month ago

MeetingRoom (Tanks and Temples - TnT)

training

python train.py -s TNT_GOF/TrainingSet/Meetingroom -m exp_TNT/Meetingroom -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m exp_TNT/Meetingroom --iteration 30000

rebecca-lay3rs commented 1 month ago

Ignatius (Tanks and Temples - TnT)

training

python train.py -s TNT_GOF/TrainingSet/Ignatius -m exp_TNT/Ignatius -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m exp_TNT/Ignatius --iteration 30000

rebecca-lay3rs commented 1 month ago

bww_entrance (nerfstudio_dataset)

training

python train.py -s data/bww_entrance -m outputs/bww_entrance -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m outputs/bww_entrance --iteration 30000

rebecca-lay3rs commented 1 month ago

building_x_1 (PPR alligned)

training

python train.py -s data/building_x_1 -m outputs/building_x_1 -r 2 --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m outputs/building_x_1 --iteration 30000

rebecca-lay3rs commented 1 month ago

building_x_1_COLMAP (COLMAP alligned)

allignement

python convert.py -s data/building_x_1_COLMAP/

training

python train.py -s data/building_x_1_COLMAP -m outputs/building_x_1_COLMAP --use_decoupled_appearance

extract the mesh after training

python extract_mesh.py -m outputs/building_x_1_COLMAP --iteration 30000