Deep Thermal Imaging: Proximate Material Type Recognition in the Wild through Deep Learning of Spatial Surface Temperature Patterns
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 14401
- Type
- E - Conference contribution
- DOI
-
10.1145/3173574.3173576
- Title of conference / published proceedings
- CHI '18: Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems
- First page
- 1
- Volume
- 2018-April
- Issue
- -
- ISSN
- 1062-9432
- Open access status
- Technical exception
- Month of publication
- April
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
3
- Research group(s)
-
-
- Citation count
- 7
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper overcomes a fundamental issue of state-of-the-art vision-based methods in in-the-wild material classification by proposing a novel machine learning-enabled thermal imaging approach. This work, thoroughly evaluated with studies conducted across indoor laboratory and outdoor real-world environment in different seasons, has demonstrated its real-time use in human-computer interaction applications. Its released datasets (over 40k images from 32 materials), software libraries have been used by over 20 institutions (e.g. Rice University, SZTAKI). Also, this work has contributed to the creation of a £5.4m EPSRC grant (Textiles Circularity Centre) and an industry research grant (£188k NTT).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -