VisiblThermalDataSet.rar

The provided data consists of processed visible and thermal images of faces from the IRIS dataset, which is part of the OTCBVS benchmark dataset collection. These images exhibit challenging variations in pose, expression, occlusion, and illumination. The dataset comprises 20 classes, with each class containing 60 samples. Prior to use, the images underwent specific preprocessing steps, including conversion to grayscale intensity, resizing to a resolution of 152x145 (originally 320x240) to enhance memory and processing speed, and cropping to isolate the facial region.