How to preprocess images for inference with MoGe v2
Nobody has claimed this yet.
- Dominant language
- JavaScript
- Stars
- 3k
- Forks
- 234
- PR merge metrics
- No merged PRs in 30d
Description
@EasternJournalist I'm trying to use MoGe v2 for inference on my own autonomous driving dataset. I have the intrinsics for all the data. However, for most of my data, the camera's principal point is not at the center of the image. For a small portion of the data, f_x is not equal to f_y.
My question is, during inference, do I need to perform the "perspective crop" mentioned here?
If I do apply this crop, how can I revert the predicted depth map back to the original image's dimensions?
Any help would be greatly appreciated!
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with moge/test/dataloader.py around line 101 and trace the inference path for camera intrinsics and image dimensions. Document whether perspective cropping is required for off-center principal points or unequal focal lengths, and how predictions map back to the original dimensions; done when the inference guidance is unambiguous.
Written by the indexing model from the issue text.
Assessment
- Domain
- computer-vision
- Issue type
- Documentation
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100