YEREVAN, Armenia — On November 12-14, the Zaven P. and Sonia Akian College of Science and Engineering (ACSE) at the American University of Armenia (AUA) presented six research projects at the 15th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), an IEEE (Institute of Electrical and Electronics Engineers) sponsored event, which was held in Barcelona, Spain. The conference brought together researchers, practitioners, and industry experts working in the field of hyperspectral data processing and remote sensing. AUA researchers, in collaboration with colleagues at the Orbeli Institute of Physiology, made a significant impact at the international meeting. They presented various aspects of their pioneering work in the hyperspectral imaging field. Specifically, they discussed the developments and applications of Multi-Excitation Hyperspectral Imaging (ME-HSI), a novel approach utilizing combinations of different excitation wavelengths.
The first study, titled “Deep Learning Classification of Ablated Atrial Tissue in Multi-Excitation Hyperspectral Images,” was authored by Arpi Hunanyan (BSDS ’22, MSCIS ’24); Nazeli Ter-Petrosyan (BSCS ’25); Fernando Villarruel and Tigran Soghomonyan, both from the L.A. Orbeli Institute of Physiology; Dr. Narine Sarvazyan, the William Frazer Endowed Professor at ACSE and a holder of the European Research Area Chair; Aram Butavyan, lecturer in ACSE; and Dr. Varduhi Yeghiazaryan, associate professor in ACSE. The last two authors are the co-principal investigators on the recently awarded Afeyan Family Foundation research grant titled “Analysis of Pre-Clinical Multidimensional Hyperspectral Datasets Using Deep Learning.” The study was presented as an oral talk by first author ArpiHunanyan. The work applies ME-HSI and deep learning techniques to improve the identification of ablated cardiac tissue, demonstrating more accurate classification compared to traditional single-excitation imaging approaches.
The second paper, co-authored by the same team, was presented as a poster and was titled “Comparative Analysis of Multi-Excitation Hyperspectral Image Configurations for Ablated Atrial Tissue Classification”. Reflecting on the experience, Hunanyan noted: “Unlike traditional HSI, which relies on a single excitation source, ME-HSI uses multiple excitation wavelengths and their corresponding emission spectra, enabling the capture of complementary biochemical features. Using ME-HSI cardiac tissue data, we classified ablated vs. non-ablated atrial tissue with deep learning models, including CNNs (Convolutional Neural Networks), Transformers, GNNs (Graph Neural Networks), and classical MLPs (Multilayer Perceptrons). We further analyzed and compared different data configurations and experimental setups to assess their impact on classification performance.”
The third study, titled “Information Detection Pipeline for 4D Hyperspectral Imaging: An Autoencoder-Based Framework with Perturbation Analysis of Complex Datasets,” was co-authored by Narek Meloyan (BSDS ’25, MSCIS ’26), Narek Chilingaryan of the L.A. Orbeli Institute of Physiology, and Dr. Sarvazyan. The research addressed the challenge of analyzing large and complex datasets generated by multi-excitation fluorescence hyperspectral imaging by applying deep learning-based autoencoder models to identify the most informative excitation-emission combinations, reducing data volume while preserving essential information and enabling more efficient and scalable biomedical imaging analysis. Meloyan, who delivered an oral presentation, remarked: “Being at WHISPERS showed me that the Armenian research community, and institutions like AUA, have the capacity to contribute meaningfully to the development of emerging technologies. With continued focus, this work can lead to even greater impact.”
The fourth study, titled “Multi-Excitation Hyperspectral Imaging Toward Improved Lichen Identification,” was co-authored by Chilingaryan, Arsen Gasparyan, assistant professor in ASCE and biodiversity researchers at the AUA Acopian Center for the Environment, and Dr. Sarvazyan. In this work, the team explored the use of the ME-HSI framework for the identification of lichen specimens. The findings highlighted the potential of ME-HSI for educational activities, citizen science, and biodiversity monitoring.
The fifth study, titled “Detection of Crosslinking-Dependent Spectral Changes in Collagen Scaffolds Using Multi-Excitation Hyperspectral Imaging,” was co-authored by Kristina Ghahramanyan (BSDS ’25); Villarruel, Chilingaryan, and Ekaterina Baidiuk, from the L.A. Orbeli Institute of Physiology; Ani Avetisyan, ACSE research assistant; and Dr. Sarvazyan. The presented data positioned ME-HSI as a powerful tool for the non-destructive, spatially-resolved quality control of collagen-based biomaterials.
The sixth study, titled “Illumination Dependence of Unmixing and Classification Performance in Multi-Excitation Hyperspectral Imaging: A Pigment Study,” was co-authored by Villarruel, Chilingaryan, and Dr. Sarvazyan. This work explored the effect that illumination quality has on the performance of diverse analysis methods applied to ME-HSI datasets.
This strong showing by AUA and the Orbeli Institute of Physiology at IEEE WHISPERS 2025 highlights the College’s growing research activities and underscores their increasing impact on global scholarship in the field of hyperspectral imaging and its applications. Three of the presented studies will be published as peer-reviewed papers in the official conference proceedings, while the others will be submitted to additional publishing venues.
Founded in 1991, the American University of Armenia (AUA) is a private, independent university located in Yerevan, Armenia, affiliated with the University of California, and accredited by the WASC Senior College and University Commission in the United States. AUA provides local and international students with Western-style education through top-quality undergraduate and graduate degree and certificate programs, promotes research and innovation, encourages civic engagement and community service, and fosters democratic values.
