In this lab, we will experiment with 3D Structure-from-Motion systems, both feature-based and feed forward-based. More specifically, you will compare the performance of a state-of-the-art feature-based approach (Colmap) with a state-of-the-art feed forward-based approach (Depth Anything 3 (DA3)).
Pull the assignment template from here.
You can download required the data directly from the notebook or from https://github.com/tsattler/zurich_graffiti_dataset/raw/refs/heads/main/images.zip
For this lab all explanations are contained in the corresponding notebook. Please, refer to the notebook for details.
To fulfil this assignment, you need to submit these files (all packed in one .zip file) into the upload system:
results.html - Converted notebook. Transform your notebook 3D_reconstruction_empty.ipynb into file results.html using jupyter nbconvert -–to html 3D_reconstruction_empty.ipynb -–EmbedImagesPreprocessor.resize=small -–output results.html
Your code and notebook will be checked manually.