OpenCV 4 with Python Blueprints - Second Edition

Understanding a perceptron

A perceptron is a binary classifier that was invented in the 1950s by Frank Rosenblatt. A perceptron calculates a weighted sum of its inputs, and, if this sum exceeds a threshold, it outputs a 1; else, it outputs a 0.

In some sense, a perceptron is integrating evidence that its afferents signal the presence (or absence) of some object instance, and if this evidence is strong enough, the perceptron will be active (or silent). This is loosely connected to what researchers believe biological neurons are doing (or can be used to do) in the brain, hence the term ANN.

A sketch of a perceptron is depicted in the following screenshot:

Here, a perceptron computes a weighted (w_i) sum of all its inputs (x_i), combined with ...