Cheng Dai
I work on thermal infrared hyperspectral imaging, HADAR-based physical vision, and efficient AI systems that move visual perception beyond appearance-only sensing. My current research focuses on temperature-emissivity-texture decomposition, physically interpretable thermal perception, and lightweight deployment for practical sensing systems.
I am actively seeking 27 Fall PhD opportunities!
I am deeply grateful for the guidance and support I have received from my mentors, senior students, and friends.
I am always open to collaborations and discussions. Please feel free to contact me at daicheng@westlake.edu.cn.
Selected Works
TeX-1500: A Paired Real-World LWIR Hyperspectral Dataset and Benchmark
TeX-1500 introduces a paired real-world long-wave infrared hyperspectral dataset for supervised HSI-to-TeX learning, together with TeX-UNet as a wavelength-aware baseline for physical-property-centered thermal perception.
Paper / Code / Dataset / Model Weights
Timeline
TeX-1500 released as a paired real-world LWIR HSI-TeX dataset and benchmark.
HAIR released as a HADAR-based restoration framework for thermal infrared hyperspectral images.
Research Assistant, AI for Physics Lab, Westlake University. Worked on HADAR algorithms, TeX decomposition, and lightweight thermal perception under the guidance of Xin Yuan and Fanglin Bao.
Audio AI Intern, Amoon AI. Worked on low-resource infant-cry activity detection and state classification, audio foundation-model fine-tuning, and edge deployment.
AI Engineering Intern, Lightwheel AI. Built engineering tools for embodied AI workflows, including RoboCasa plugin development and teleoperation tooling.
Graduate Study, Wuhan University. Studied Communication Engineering at the State Key Laboratory of Surveying, Mapping and Remote Sensing Information Engineering.
B.Eng. in Communication Engineering, Jilin University. Built foundations in signal processing, communication systems, mathematical modeling, and engineering practice.