Neural Networks (NN) is one of the most fundamental building blocks of modern Computer vision and it has found its relevance in multiple domains like image, text, audio, and many more. Various operations, convolution operation in analysing images to recurrent neural networks used to process text uses the principle of NN. Also, researchers have come up with novel deep learning architecture to analyse and find high-order patterns in domains like 3D structure.

However, it is not easy to implement domains like 3D computer vision with essential utilities of NN library (which is primarily focused on classic computer vision). Implementation can generally become complicated and confusing. Also, it's no longer deemed necessary among deep learning enthusiasts to know the traditional computer vision algorithm from the implementation point of view, but they are beneficial from the understanding point of view. Therefore a framework which can unifies and ease the task of 3D vision will be useful to the community.



Nirmal Praveen Suthar


  • Elliot Saba
  • Avik Pal
  • Dhairya Gandhi