Running Deep Learning models in OpenCV

by Ankit Sachan • July 12, 2018

OpenCV: 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. Deep […]

Continue Reading

Choosing a Deep Learning Framework in 2018: Tensorflow or Pytorch?

by Ankit Sachan • May 29, 2018

One of my friends is the founder and Chief data scientist at a very successful deep learning startup. 2017 was a good year for his startup with funding and increasing adoption. However, on a Thursday evening last year, my friend was very frustrated and disappointed. The framework on which they had built everything in last […]

Continue Reading

Breaking Deep Learning with Adversarial examples using Tensorflow

by Richeek Awasthi • May 22, 2018

It’s no news that Deep Learning is super effective and powerful in solving computer vision problems. To keep things in perspective, the top 5 accuracy of NASnet on the ImageNet dataset is 96.2% which is greater than human accuracy on the same task(approx 94.9%). Similarly, deep learning has surpassed or equaled the human level accuracy for […]

Continue Reading

Bias-Variance trade-off in Machine Learning

by Ankit Sachan • January 17, 2018

In this post, we shall talk about bias-variance trade-off in machine learning and tips and tricks to avoid overfitting/underfitting. Let’s start with a real-world example. Fukushima power plant disaster: Failure of predictive modeling What does a nuclear power plant disaster have to do with machine learning? The safety plan for Fukushima Daiichi nuclear power plant was […]

Continue Reading

Object Detection using Single Shot Multibox Detector

by koustubh • January 2, 2018

In a previous post, we covered various methods of object detection using deep learning. In this blog, I will cover Single Shot Multibox Detector in more details. SSD is one of the most popular object detection algorithms due to its ease of implementation and good accuracy vs computation required ratio. Work proposed by Christian Szegedy […]

Continue Reading