Pixel Intensity Plots

Background:

A histogram is a graphical representation of data. In the context of image processing, histograms are primarily used to visualize pixel color distribution of an image. The x-axis of a histogram will show the variety of pixel intensities. The y-axis is the count of each pixel intensity.

By visualizing the data of an image’s color planes in a histogram, we can associate a specific distribution of colors with a given category. For example we would expect an image of a red apple to have more pixels in the red channel than pixels in the blue channel, so if we have a histogram with a larger red intensity then we know our image is more likely to be an apple.

We use histograms to understand how a machine learning model might work. Similar to histograms, a machine learning model can use a dataset of a specific category with many different images and then compare the histograms of these images to find patterns or similarities. These comparisons can be used to more accurately predict a given input.