Supplement of paper "Making Infrared Face Recognition Robust Against Perceptual Image Quality Degradation", and LWIR facial datasets files.

These files are supplementary to the research projects involving LWIR face recognition and thermal-visual spectrum fusion carried in the Pontificia Universidad Javeriana at Cali, Colombia.<br>The 'supplement.pdf' document contains supplementary results of the study on image quality assessment applied on infrared face recognition systems, described in the paper "Making Infrared Face Recognition Robust Against Perceptual Image Quality Degradation".<br>The ZIP files correspond to the two datasets of Long Wave Infrared (LWIR) face images created at the Pontificia Universidad Javeriana at Cali, Colombia from 2015 to 2016. The PUJ-T360 collection was used in a study on the influence of image quality distortions on thermal face recognition and application of Image Quality Assessment techniques for enhancing a face recognition algorithm (http://bit.ly/IQA_IR_FR).<br>Both data sets comprises aproximately 800 LWIR and visual-spectrum images from 40 subjects (10 thermal and 10 visible images per subject). The images were acquired with two different illumination conditions (lights on and off), three different poses (frontal, left and right profiles) and three different expressions (neutral, smiling and surprised). Additional images with eyeglasses were included for a few subjects. LWIR and visual-spectrum images in PUJ-FONE are co-registered (aligned); images in PUJ-T360 are not.<br>The PUJ-T360 collection was acquired using a FLIR T360 camera, and for the PUJ-FONE we used a FLIR ONE (2nd. generation) camera, attached to an iOS device (Apple iPhone or iPad).