TeX-1500: A Paired Real-World LWIR Hyperspectral Dataset and Benchmark for Temperature-Emissivity-Texture Decomposition
Researchers have introduced TeX-1500, a new dataset designed to advance temperature-emissivity-texture (TeX) decomposition from long-wave infrared hyperspectral imaging (LWIR HSI). This dataset comprises over 1,500 paired real-world scenes, bridging the gap in supervised learning for TeX decomposition. It includes calibrated radiance cubes, wavelength positions, and aligned temperature, emissivity, and texture data, along with a baseline model called TeX-UNet. AI
IMPACT Enables more robust learning-based decomposition of thermal and material properties from hyperspectral imagery.