TY - GEN
T1 - Comparison of Hyperspectral Image Reconstruction for Medical Images
AU - Mohammed Ridha, Ali
AU - Isa, Nor Ashidi Mat
AU - Tawfik, Ayman
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
PY - 2026
Y1 - 2026
N2 - Hyperspectral imaging (HSI) which captures a wide spectrum of light has emerged as a tool for the detection and diagnosis of various medical conditions. However, due to the high cost of specialized HS cameras, it is limited in its use in clinical settings. In this research, a comprehensive comparison is carried out between two architectures for hyperspectral reconstruction algorithms for medical images of acne vulgaris. The evaluation will consist of an analysis of different hyperparameter configurations to identify the optimal reconstruction algorithm for medical hyperspectral images. The results show that the HRNET architecture model, which includes colour correction, random cropping, and a small batch size had the lowest mean relative absolute error of 0.0433. Therefore, the reconstructed hyperspectral (HS) images using HRNET architecture could offer a viable and cost-effective alternative to utilizing expensive hyperspectral imaging (HSI) equipment for detecting medical conditions.
AB - Hyperspectral imaging (HSI) which captures a wide spectrum of light has emerged as a tool for the detection and diagnosis of various medical conditions. However, due to the high cost of specialized HS cameras, it is limited in its use in clinical settings. In this research, a comprehensive comparison is carried out between two architectures for hyperspectral reconstruction algorithms for medical images of acne vulgaris. The evaluation will consist of an analysis of different hyperparameter configurations to identify the optimal reconstruction algorithm for medical hyperspectral images. The results show that the HRNET architecture model, which includes colour correction, random cropping, and a small batch size had the lowest mean relative absolute error of 0.0433. Therefore, the reconstructed hyperspectral (HS) images using HRNET architecture could offer a viable and cost-effective alternative to utilizing expensive hyperspectral imaging (HSI) equipment for detecting medical conditions.
KW - Hyperspectral Imaging
KW - Hyperspectral Reconstruction
KW - Machine Learning
UR - https://www.scopus.com/pages/publications/105019041430
U2 - 10.1007/978-3-031-98498-3_2
DO - 10.1007/978-3-031-98498-3_2
M3 - Conference contribution
AN - SCOPUS:105019041430
SN - 9783031984976
T3 - Communications in Computer and Information Science
SP - 18
EP - 32
BT - Applications of Artificial Intelligence and Data Science - 1st Global Conference, AAIDS 2024, Proceedings
A2 - Mahmud, Mufti
A2 - Pillay, Nelishia
A2 - Kaiser, M Shamim
PB - Springer Science and Business Media Deutschland GmbH
T2 - 1st Global Conference on Applications of Artificial Intelligence and Data Science, AAIDS 2024
Y2 - 3 April 2024 through 5 April 2024
ER -