You can probably figure out the image above — pretty easy, right! What if this task was assigned to a machine or your computer,will it still be an easy task.It’s not that simple for computers if you are wondering, as they do not the gift of vision and inbuilt human-like classification prowess. So it becomes one of the preliminary task to communicate with your machine which is obviously in the form of 1’s and 0’s , and if anyone has prior knowledge that an image can be represented in a form of binary digits which makes the communicating easier. With the recent advances in GPU capacity and convolutional neural network, computer vision has gained great acclaim. Computers now give us better accuracy than humans for images that are quite complex and have features that are not easily differentiable by naked eye, but then again computers have the prowess to figure out the details stringently. Let’s take a dive into the endless scope of machine learning by just exploring a sub-domain of its many application — classifying images.

Introduction:

For the keeping things simple, we are going to use Logistic Regression for image classification. The data set for our study is one of the most popular handwritten digits know as MNIST dataset. In other words, given an image of a handwritten digit, we want to classify it as a 0, 1, 2, 3, … . Before we get started let’s import the necessary libraries.

#data-analysis #machine-learning #data-science