PIX-IT

Team PIX-IT has worked on the Fast Bilateral-Space Stereo for Synthetic Defocus paper by researchers at Google.

Requirements

This project expects the following libraries to be present on the users system along with Python3

These libraries can be installed via pip

• Jupyter: tools for working with code, text in one place
• OpenCV: a library of programming functions mainly aimed at real-time computer vision
• Numpy: for dealing with large, multi-dimensional arrays and matrices
• MatplotLib: a plotting library
• SciPy: for scientific computing
• PyLab: an interface to Matplotlib (Generally gets installed with MatPlotLib by default)

To run the code!

• Please open the 'src' folder
• The 'BilateralSolver.ipynb' is our fully functional notebook.
  • It includes all the modules, classes
  • It runs all the tests
• Simply run all cells of 'BilateralSolver.ipynb' in order.

Documentation of the code

• Check out index.html in ./references/documentation for the documentation of the code.
• Documentation created using Sphinx documentation generator.

Sources for stereo images

• https://vision.middlebury.edu/stereo/data/
• https://vision.deis.unibo.it/fede/ds-stereo-lab.html

Repository Structure

• documents
  • PIX-IT-Final.pdf: the PDF version of final presentation. (Recommended for presentation!)
  • PIX-IT-Final.pptx
  • PIX-IT_mideval_presentation.pdf
  • PIX-IT_mideval_presentation.pptx
• images: contains images used in code (4 pair of stereo images)
• references: contains reference papers
  • documentation_html: open index.html file to run the documentation on the browser
• src
  • own: [ARCHIVE] contains rough notebooks where we tried and tested different approaches
  • BilateralStereo.ipynb: Main Notebook. Run this!
• readme.md: [This File]
• proposal.md: project proposal