Project DetailsThermal image colorization is a topic that is gaining momentum in the world of artificial intelligence. In recent years, with a significant improvement in the tools, and with the algorithmic development of deep learning, the world of computer vision has managed to achieve impressive achievements in everything related to image processing and analysis.
A significant development that has led to the rapid increase in achievements is the development of the CNN called GAN - Generative Adversarial Network. Networks of this type make it possible to produce a new set of information based on the characteristics of the existing information.
In the current project, the Cycle-Gan approach was tested for the problem of domain transfer from TIR to RGB. The projects goal was to achieve the best results, both metrically and visually, for the given problem, while acquiring data using a flir-1 portable thermal camera. The network was trained and adapted to the needs of the given goal, and the effect of various loss functions on the result was examined. While using Cycle-Gan two main methods were examined: aligned images and unaligned images. The results for the unaligned images were better.
