Week 7: Legible or not? (Part 2)

Apr 07, 2019

During this week, I primarily focused on classifying the numbers from the interesting handwriting sample that Ms. Bhattacharya gave me. After trying numerous packages and deciding on OpenImageR last week, I started playing around with it this week. First, I gave the program the path to the image I stored on my laptop. Then, using the OpenImageR package, I used the readImage function to load the image into R. This function works in a way such that each image has 3 copies – one with the red values, one with the blue values, and one with the green values. Since my neural network only takes in grayscale images, I changed this image from RGB to grayscale using “rgb_2gray.”  After this, I reshaped the image to the size of the previous images I had been using and then had my neural network predict what number the image was of. I had a couple of successful runs and some unsuccessful ones.

Classified as a 7

Classified as a 1

 

Next week, I hope to figure out the issues with this classification and why it read the 5 as a 1.

Best,

Arshia

One Reply to “Week 7: Legible or not? (Part 2)”

  1. Vidur Gupta says:

    How many images did you have in your training set?

Leave a Reply

Your email address will not be published.