Elimination of defects in photos using neural networks

Abstract

The use of the mathematical apparatus of neural networks is becoming a very important tool in solving various problems of artificial intelligence. Pattern recognition problems can be attributed to the number of such problems. This paper provides an example of using a neural network to identify a defect in a photograph, and then eliminate such a defect.

Materials and methods, results and discussions. The proposed article discusses special algorithms for the operation of a neural network that are designed to recognize graphic images, as well as algorithms for finding defects in graphic images with their subsequent elimination.

A huge number of photographs on paper are subject to physical damage over time (unlike photographs on electronic media). To solve such a problem, it is necessary to develop algorithms for searching and recognizing defects in photographs, as well as algorithms for eliminating these defects using the mathematical apparatus of a neural network.

The article discusses approaches to the development of a neural network for recognizing defects in photographs with their subsequent restoration.

Conclusion. The solutions that are offered based on the results of the work performed within the framework of this article will enable users to restore old photographs with defects that can be invaluable not only for them, but for the whole society - for example, photographs of historical events, great people of our state.