/gpucv

A easy-to-use image processing library accelerated with CUDA on GPU.

Primary LanguageJupyter NotebookMIT LicenseMIT

gpucv

Open In Colab

Have you used OpenCV on your CPU, and wanted to run it on GPU. Did you try installing OpenCV and get frustrated with its installation. Fret not gpucv is here to save the day. Simple installation and runs as good as OpenCV on GPU.

You need to still install all the Nvidia drivers if you don't have them.

Requirements

  • Python 3.7
  • CUDA

Installation

  1. Using pip
pip3 install gpucv
  1. Building the project

Clone the project to your local repo and run setup.py.

git clone git@github.com:shrikumaran/gpucv.git
python3 setup.py bdist_wheel

Example

import gpucv
img = gpucv.readimg('index.jpeg')
sobel = gpucv.sobel(img)

original sobel

Colab noteobook: https://colab.research.google.com/drive/1o0LD56Qo88lZZtfVqshZL-rvCZUJ19ex?usp=sharing

TODO

  • Basic IP stuff (grayscale,resize,threshold)
  • Implement filters (mean,gaussian and sobel)
  • Convolution kernels