microsoft / microsoft/MoGe

How to train with GT pointmap instead of Depth Images

Open
#63 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
JavaScript
Stars
3k
Forks
234
PR merge metrics
No merged PRs in 30d

Description

Hey there, thanks for this awesome project. I have been able to generate some pointmaps, i.e. actual (x, y, z) per pixel, and was wondering if there was a way to train MoGe with that as the ground truth instead of 16-bit depth images. It looks like I would need to rewrite read_depth in io.py, and alter the dataloader as well for this, but was wondering if there was a better way. Thanks for your help

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading read_depth in io.py and the dataloader path mentioned in the issue to understand how 16-bit depth targets enter training. Clarify the expected pointmap representation and training target interface; done means establishing a supported path for using per-pixel (x, y, z) ground truth instead of depth images.

Written by the indexing model from the issue text.

Assessment

Domain
computer-vision, machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.