MWIR Thermal Image Dataset for Deep Learning Based Super Resolution

This dataset contains 191 real Mid-Wave Infrared (MWIR) thermal images acquired using a cooled infrared camera. The images have a spatial resolution of 640 × 512 pixels. The dataset is divided into 153 training images and 38 validation images, corresponding to the split used in the associated super-resolution study. For the experiments, ×4 low-resolution inputs were generated from the original high-resolution images using bicubic downsampling. The dataset is intended to support research on thermal imaging, image restoration, and deep learning-based super-resolution.