TeX-1500: A Paired Real-World LWIR Hyperspectral Dataset and Benchmark for Temperature-Emissivity-Texture Decomposition

Temperature-emissivity-texture (TeX) decomposition aims to recover object heat state, material spectral response, and visible-like geometric texture from long-wave infrared hyperspectral imaging (LWIR HSI). Learning-based TeX decomposition has been limited by the lack of paired real-world LWIR HSI-TeX supervision.

This work introduces TeX-1500, a paired real-world dataset and benchmark for supervised HSI-to-TeX decomposition. The dataset contains 1,522 calibrated real-scene pairs from DARPA Invisible Headlights pushbroom imagery and FTIR acquisitions, spanning multiple locations, seasons, acquisition times, wavelength layouts, and sensor families.

The paper also provides TeX-UNet, a simple wavelength-aware baseline that maps calibrated HSI bands and wavelength positions to TeX fields. Experiments on held-out DARPA IH scenes and zero-/few-shot transfer to FTIR scenes show that TeX-1500 supports measurable progress toward data-driven, physical-property-centered thermal perception.

Recommended citation: Cheng Dai, Jiale Lin, Hongyi Xu, Bingxuan Song, Ziyang Xie, and Fanglin Bao. (2026). "TeX-1500: A Paired Real-World LWIR Hyperspectral Dataset and Benchmark for Temperature-Emissivity-Texture Decomposition." arXiv preprint arXiv:2606.03806.
Download Paper