What are the advantages of a Hamming neural network over a Hamming distance comparison algorithm?

Question

Neural networks

Task: to say which of the given images (or patterns) is most similar to the input image (or input pattern).
I'm being asked to solve this using a Hamming Neural Network, but I think it can be done in a simpler way - just write an algorithm that will calculate the Hamming distance between a given image and images in the database. That, in comparison with which the Hamming distance will be minimal, will be the most similar.
What are the downsides of my approach compared to using a Hamming neural network?

Answer 1

Actually, according to Wikipedia
So what you are contrasting with is not entirely clear.
But here's something else to worry about. It seems to me that the concept of "most similar image" is very vague. Well, for example, does it imply the similarity of pictures, the image on which is of different scales? Or does he use different colors? Or contains graphical image distortion? Or simply - are these images that depict (no matter how) objects of the same class? Etc.
And it seems to me that the Hamming Neural Network cannot cope with most of these tasks by definition.