Skip to content

Latest commit

 

History

History
24 lines (16 loc) · 851 Bytes

README.md

File metadata and controls

24 lines (16 loc) · 851 Bytes

3D Reconstruction

Description

This repository implements 3D Reconstruction (from RGB images) using Structure from Motion Pipeline. The pipeline consists of

  1. Feature Detection (using SURF / SIFT / Hessian Affine Region Detector)
  2. Feature Matching (using FLANN)
  3. Reconstruction (using Patch Match)

Pipeline

alt text

Results

Heres the result of SfM on Flash Action Figure

alt text

Closing Thoughts

Although we get decent reconstruction of the figure, it can be further improved

  1. Statistical or Radial Outlier Removal can further help in noise removal from the dense reconstruction.
  2. The head in front view has not been reconstructed inspite of many images in same pose. This is due to glare from the light source above it. Image processing algorithms can help fix this issue.