/greyscale-detection

Primary LanguageJupyter Notebook

Grayscale-Detection

Pure Computer Vision Method

Tools used:

  • OpenCV
  • Numpy
  • Matplotlib

Explaination:

I have divided image into 3 channels (r,g,b) and have taken the sum of the differences between each channel. If the given image belongs to the grayscale category, sum will be zero. If not, it contains some amount of colour. Next, we calculate percentage of grayscale. This is done by finding the ratio of the sum (calculated above) to the image size. This will give is the colour percentage, hence, for grayscale we subtract it from one.

Convolutional Neural Network Classification

Tools used:

  • Keras
  • Numpy

Explanation

Here I have used a small custom dataset using google to simply classify if it's a coloured image or a grayscale image. Here we ahve used a 3 layer convolutional neural network to train the model. Test cases are given below.