Help Needed - Resolving .ply Quality Issues in MonoGS 3D Reconstruction with Custom Data

[NagaPadma]

Hello SLAM Community,

I am currently working on a 3D scene reconstruction project using the MonoGS repository and could use some advice. I have a video sequence of 374 images and have prepared the necessary data files including a groundtruth.txt from COLMAP and depth images from Depth Anything V2.0. I’ve also updated the configuration file with the calibration matrix generated from COLMAP.

Despite following the guidelines and ensuring the data is correctly formatted and manipulated, I am experiencing issues when running slam.py.

To avoid previously reported issues, I have commented out the following code:
eval_ate(
self.cameras,
self.kf_indices,
self.save_dir,
0,
final=True,
monocular=self.monocular,
)
However, the resulting .ply file does not meet the expected benchmarks. I am eager for any fixes or insights that could address this issue. Debugging at this stage has been quite challenging, and any shared experiences or advice on maintaining output quality while resolving related errors would be greatly appreciated.

Thank you for any help you can provide! Looking forward.