| Abstract [eng] |
Color change in pictures using artificial neural networks is about using artificial neural networks to change grayscale images to coloured images. This thesis is part of a larger part of computer imagery and imagery analysis. Artificial neural networks are used in a very wide variety of tasks. As technology progresses, so does artificial neural networks and efficiency in which they are doing their tasks. There are multiple ways how artificial neural networks can be used to colour images. They differ in their processes, how they work, results they show and others. There are also problems which exist by using neural networks to colour images. In this thesis, an artificial neural network was chosen, to analyze its capabilities and limitations. Also, different parameters of this neural network were analyzed, as well as different architectures to get as good as possible results. Final coloured images were compared using image comparison algorithms. Best results were compared with other artificial neural networks. |
