Night Vision Colorization from Color Mapping to Color Transferring

Multispectral imaging greatly supports executing nighttime object assessment tasks. In such a context, analyzing colorized multisensory images definitely improves observer situational awareness, reaction time, and perceptual analysis (human vision). Night vision (NV) colorization is such a technique that renders multispectral images (in grayscale) with daylight colors. This paper provides a brief review of NV colorization techniques that includes well-known color mapping methods that use statistical matching, and the recently-introduced color transferring methods that use a convolutional neural network (CNN) to map colors. Objective evaluation index (OEI) is used as a quantitative metric to compare colorization results. The experimental results show that the color transferring methods using CNN is very promising in colorizing NV imagery.