/3d-reconstruction_PIFu

3D Reconstruction using PIFu

Converting 2D Single Image into 3D Model

2D Single Image to 3D Model using the PIFu algorithm.

We utilized the PIFu algorithm(Saito et al. 2019) to create a 3D model of a sagittal mouse brain with a single 2D image as an input. The sagittal mouse brain image and binary masks were obtained from the source from Iqbal et al(2019).


3d_1

3d_2

3d_3

3d_4

Code

git clone https://github.com/shunsukesaito/PIFu.git

cd /content/PIFu

!sh ./scripts/download_trained_model.sh

!sh ./scripts/test.sh

!pip install 'git+https://github.com/facebookresearch/pytorch3d.git@stable'

from lib.colab_util import *
obj_path = '/content/PIFu/results/pifu_demo/result_mousebrain.obj'
video_path = '/content/PIFu/results/pifu_demo/result_mousebrain.mp4'
video_display_path = '/content/PIFu/results/pifu_demo/result_mousebrain_display.mp4'

renderer = set_renderer()
generate_video_from_obj(obj_path, video_path, renderer)

!ffmpeg -i $video_path -vcodec libx264 $video_display_path -y -loglevel quiet
video(video_display_path)