Programmatically identify which camera clicks better photos

Programmatically identify which camera clicks better photos

Credit: Ritesh Bhagwat

Let's say we have 3 cameras. Canon, Nikon, and Sony and we take pictures of Sunset on a beach. All three cameras have slightly different software processing so the output of all three cameras will be somewhat different.

Now the question is which is the best picture? So the 'truth' here is finding out which is the best picture. Well, it depends on personal taste. If someone likes high contrast then a certain image is the best image, if someone likes more vibrant colors another picture is the best picture.

The keyword here is 'Subjective'. The best picture is totally subjective and dependent on the person's taste. Many of you may disagree with me but I feel on a broader 'Spiritual' level truth is also Subjective.

Probability is the best tool to study subjectivity. There are three types of Probability.

  1. Theoretical Probability.
  2. Empirical Probability (Loosely referred to as Frequentist Probability)
  3. Subjective Probability (Loosely referred to as Bayesian Probability)

To study Subjectivity, the Bayesian approach to probability is very powerful.

As we say truth is subjective if you show the picture to a dog, the dog will laugh at you as he sees the world only in black and white. The perception of colors is absent in a dog so his truth is different than our truth!