Machine Learning Optimization for Colour Image Reconstruction from Thermal/Infrared Images

Abstract

Machine Learning has generated a tremendous interest in research and development under the umbrella of Artificial Intelligence. It was a field that evolved from pattern recognition and computational learning in Artificial Intelligence. Machine Learning algorithms are capable of identifying how to accomplish certain tasks by generalizing from real world examples. In comparison to manual programming, this is often feasible and cost-effective. In this context, this paper focuses on the use of these optimized machine learning algorithms in order to reconstruct colour images from thermal/infrared images. With the capability of these machine learning algorithms to identify patterns in existing data sets, such an algorithm can be used to reconstruct a colour images based on the features that are given to the algorithm. Most night time photography uses thermal imagery because of its capability of capturing thermal radiation from the human body. So these data can be used to reconstruct a colour image based in feature recognition form the thermal image. In conclusively it is our intention to use the power of machine learning techniques to build a system that can generate colour images by analysing the small amount of details that are present in thermal images.