Tasking on natural statistics of infrared images

Natural Scene Statistics (NSS) provide powerful perceptually relevant tools that have been successfully used for image quality analysis of visible light images. NSS capture statistical regularities that arise in the physical world and thus are relevant to Long Wave Infrared (LWIR) images. LWIR images are similar to visible light images and mainly differ by the wavelengths captured by the sensors. The distortions unique to LWIR are of particular interest to current researchers. We analyze a few common LWIR distortions and how they relate to NSS models. Humans are the most important factor for assessing distortion and quality in IR images, which are often used in perceptual tasks. Therefore, predicting human performance when a task involving LWIR images needs to be performed can be critical to improving task efficacy. The National Institute for Standards and Technology (NIST) characterizes human Targeting Task Performance (TTP) by asking firefighters to identify the locations of fire hazards in LWIR images under distorted conditions. We find that task performance can be predicted using NSS features. We also report the results of a human study. We analyzed the NSS of LWIR images under pristine and distorted conditions using four databases of LWIR images. Each database was captured with a different camera allowing us to better evaluate the statistics of LWIR images independent of camera model. We find that models of NSS are also effective for measuring distortions in the presence of other independent distortions.