Shape Analysis 101

To use

1. Create conda environment from yaml file:

   conda env create -f shapeanalysis101_env.yml
   

2. Activate environment:

   source activate shapeanalysis101
   

3. Start jupyter server:

   jupyter notebook
   

4. Explore and modify the various .ipynb notebooks

5. Answer sheets can be found on the solutions folder. To run them, simply move the [Solution] notebook you would like to run to the base shapeanalysis101 repository.

Troubleshooting

If the shapeanalysis101_env.yml fails to install for whichever reason, try and use env_wo_gpu.yml instead as follows:

conda env create -f env_wo_gpu.yml
source activate shapeanalysis101_2

If env_wo_gpu.yml also fails to install, you can create a fresh environment that will be sufficient to run notebook 1, 2A, and 2B with the following commands:

conda create -y -n shapeanalysis101_3 -c conda-forge python=3.8
source activate shapeanalysis101_3
conda install jupyter
conda install h5py
conda install imageio
conda install matplotlib
conda install scikit-learn
conda install scikit-image

You can then resume at step 3 above.

To cite

If you use the exercise notebooks from this tutorial, please acknowledge it as follows:

Hugger, J., Uhlmann, V. (2021). Shape Analysis 101 Tutorial. github.com/uhlmanngroup/shapeanalysis101

If you reuse pieces of code from this course for your own research, please acknowledge them as follows:

• For the spline models library spline_curve_model:

Uhlmann group's Spline Fitting Toolbox v1.0

• For 2B - Continuous Shape Analysis:

Song, A., Uhlmann, V., Fageot, J., & Unser, M. (2020). Dictionary learning for two-dimensional Kendall shapes. SIAM Journal on Imaging Sciences, 13(1), 141-175.

• For 3 - Shape Embeddings:

Hugger, J., Uhlmann, V., (2021). Shape Embeddings for Biological Morphology Quantification. Preprint.

Further reading

• Dryden, I. L., & Mardia, K. V. (2016). Statistical shape analysis: with applications in R. John Wiley & Sons.
• Klingenberg, C. P. (2020). Walking on Kendall’s Shape Space: Understanding Shape Spaces and Their Coordinate Systems. Evolutionary Biology, 47(4), 334–352.
• Khosla, P., Teterwak, P., Wang, C., Sarna, A., Tian, Y., Isola, P., Maschinot, A., Liu, C. & Krishnan, D. (2020). Supervised contrastive learning. arXiv:2004.11362.

Acknowledgements

Part of the material from this tutorial was adapted from Paula Balcells' Bachelor Thesis work carried out in the Uhlmann group. We also thank Anna Song for useful discussions and Jean Feydy for sharing his excellent teaching material (https://www.jeanfeydy.com/Teaching/), which inspired part of this course's structure.