tensorflow / tensorflow/graphics

NASA: generating input tfrecord files

Open
#467 0 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
2.8k
Forks
374
PR merge metrics
No merged PRs in 30d

Description

Hi,
I have some questions about generating input dataset for NASA.

Could you open the script used when generating the sample tfrecord data?

According to section 6.1 in the paper, training data consists of two:

  1. randomly sample points uniformly within a bounding box scaled to 110% of its original diagonal dimension
  2. Poisson disk sampling on the surface with normal noise (sigma 0.03)
    and the indicator labels are computed by randomized ray casting and checking the parity

I'm wondering how this sampling process was implemented.
Thank you for your great work !

Contributor guide

Open the contributing guide

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

Review tensorflow_graphics/projects/nasa and the “Prepare the dataset” section linked in the issue, then identify where the sample TFRecord data generation is implemented. Document or expose the process for uniform bounding-box sampling, surface Poisson-disk sampling with normal noise, and randomized ray-casting labels, so users can reproduce the sample data.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, tensorflow
Domain
data, machine-learning
Issue type
Documentation
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.