The largest computer vision library OpenCV can now deploy Deep learning models from various frameworks such as Tensorflow, Caffe, Darknet, Torch.

A. The open source computer vision library for everyone:

OpenCV has been the go-to library for computer vision for more than a decade. Released under BSD license, OpenCV code is free to be used for academic as well as commercial purposes. Millions of developers/academics/companies have been using OpenCV to build some really cool stuff.

B. Deep Learning changed computer vision forever:

the world of computer vision changed when deep learning arrived. For most of the computer vision tasks, deep learning models were built and trained which started outpacing the counter-part old machine learning methods implemented in OpenCV. So, accuracies got a shot in the arm. Time taken to build computer vision systems shrank to months from years(We didn’t need a complicated feature engineering process for each task). And all of this has happened too fast. Within a few years, deep learning has completely disrupted the computer vision. It has been all too exciting for a while, really hot and heavy. Even for some fundamental task like image classification, object detection etc. the state-of-the-art methods and accuracies were changing every few months.

C. Deep Learning in OpenCV:

OpenCV has introduced the flexibility of deploying deep learning models trained on other frameworks in OpenCV. Currently, it supports Caffe, Torch, Tensorflow, and Darknet. This is especially useful if you have deployed OpenCV based model say Hog+SVM classifier or Haar cascade based detector etc. in production and you want to replace it with a more accurate appropriate deep learning based model.  So, we can train your models in your favorite deep learning library and deploy that model in your production infrastructure without disrupting your current workflow.