R2-Gaussian note

R2G note

Posted by SA on January 10, 2025

Foreword

At present, the R2-Gaussian related code has been released, recording some precautions for R2-Gaussian

cover

demo

1. 2-Views Reconstruction

In dataset_readers.py line 281 and 278 modify the code


angles = np.sort(np.random.rand(50) * 360 / 180 * np.pi) + data["startAngle"] / 180 * np.pi
# angles = data_split["angles"]
        
# for i_split in range(n_split):  # default
for i_split in [0, 25]:  # 2 views

Reference