No-reference task performance prediction on distorted LWIR images

Recent work on the problem of Image Quality Assessment (IQA) has produced accurate subjective quality evaluators for visible light images. Two such algorithms are the Blind/Referenceless Image Spatial QUality Evaluator (BRISQUE) and the Natural Image Quality Evaluator (NIQE). Both models are useful in that they correlate highly with human visual perception of image quality. Given that other kinds of non-visible light images are also ´natural´ projections of the world, and can be distorted thereby reducing the perceived quality, it is of interest to study whether quality prediction on other image modality can find practical use. To this end we have extended the application of modern blind IQA models.