search

shunsukesaito / PIFu

This repository contains the code for the paper "PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitization"
https://shunsukesaito.github.io/PIFu/
Other
1.76k stars 341 forks source link

Data Generation #30

Closed zyz-notebooks closed 4 years ago

zyz-notebooks commented 4 years ago

Regarding data generation, what are the points to pay attention to in the preparation of training data, or can you recommend some data sets for learning, thank you very much

shunsukesaito commented 4 years ago

Can you elaborate more on what you want to know? What kind of applications do you intend to use PIFu for? For clothed human digitization, I'd recommend reading PIFu's supplementation material and this code carefully.

zyz-notebooks commented 4 years ago

In this aspect of furniture, such as shapenetCoreV1, is it okay? I hope to further learn 3D reconstruction by running your code. Much appreciations if your could reply my question.

shunsukesaito commented 4 years ago

The obj files in ShapeNet are non-watertight. Thus, you need to first convert them into occupancy fields. Previously when I tested on ShapeNet, I used this data. But it'd be best to follow sampling strategy in the original PIFu to achieve highly detailed geometry.

zyz-notebooks commented 4 years ago

Thank you very much for your answer, which solves my question