-
Dear author,
Thank you so much for this amazing repo!
I have question regarding training instability with certain loss combinations.
I trained two models:
(1) 1.0*L1+1.0*L2+1.0*DDIM (baseline…
-
Thanks for your great work!
But I have a question: In the sfm reconstruction script, feature extraction and matching were first performed using colmap, and then feature extraction and feature match…
-
After running the dense reconstruction in Colmap, I would like to estimate the normals at each point of the generated point cloud for downstream tasks. I'm currently doing this using the Open3D librar…
-
Create table for words and nouns. Query it when generating usernames
Reference: https://stackoverflow.com/questions/8674718/best-way-to-select-random-rows-postgresql
-
In the original paper it stated that the tetra-hedra field can be obtained from the sparse or dense point cloud.
How can we use the dense point cloud ?
-
Thanks for the great work and open-sourcing this useful tool.
It seems that the tool only needs sparse point clouds, camera poses, and images. The image distortion should be negligible for decidin…
-
How to do the mesh reconstruction and mesh texturing when **only** having access to:
- a known dense point cloud (colored, without normals),
- known camera intrinsics and extrinsics
- and the ima…
-
I try my code:
# python standard libraries
import os
import random
import fnmatch
import datetime
import pickle
# data processing
import numpy as np
np.set_printoptions(formatter={'float_ki…
-
Hello! I want to make sure I am using the depth maps generated in the dense reconstruction step correctly. Searching the docs, scripts and discussions I couldn't find a definitive specification for wh…
-
The monodepth2 considers the middle frame `t` as the target frame and makes a forward sequence warps `t-1->t, t+1->t`. Since the target is the same for the photometric error, the minimum is considered…