Skip to main content
2024 BMES Annual Meeting Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Biomedical Imaging and Instrumentation

Biomedical Imaging and Instrumentation - Poster Session C

Poster F4 - HIA-NN Enhanced Precise Subcellular Metabolic Profiling in Hyperspectral SRS Imaging

Friday, October 25, 2024
10:00 AM - 11:00 AM EST
Location: Exhibit Hall E, F & G

Introduction:  Aging and diseases, as significant drivers of mortality, are frequently linked to complex metabolic dysregulation. Metabolic dysregulation, caused by abnormal gene expression, mislocalization of molecules, and organelle dysfunction, complicates our understanding of cellular processes and disease progression. Precise and advanced analytical tools are essential for exploring these intricate metabolic pathways and uncovering the underlying mechanisms of age-related diseases and their treatments. Stimulated Raman Scattering (SRS) imaging technology and hyperspectral images, offers resolution and specificity in these research. Recognizing the big need for more efficient research methodologies from the researchers, we developed HIA-NN, a neural network program that enables rapid and precise analysis of hyperspectral images of biological samples via SRS , poised to achieve more efficient research process.

Materials and

Methods:  HIA-NN, as a machine learning driven neural network program, utilizes Stimulated Raman Scattering Imaging Technology (SRS) to generate hyperspectral images (HSI) for biological samples. This technology enhances the accuracy and efficiency of bioimaging research by allowing the automatic identification and quantification of molecular types, automatic localization of the molecules within cells, and conduct detailed intensity profiling and ratiometric comparisons. While analyze the hyperspectral image(HSI), HIA-NN targeting the -CH vibrational region (2700-3200 wavenumbers) and the CD vibrational region (2000-2450 wavenumbers) which are crucial for identifying newly synthesized organic compounds. HIA-NN refine and analyze the HSI by experiencing several steps: Noise Signal Removal and Spectral Fitting, Peak Studying, Peak Parameter Prediction, and Image Refinement. Ultimately, HIA-NN not only quantifies and identifies the types of molecules involved in metabolic processes but also provides refined hyperspectral images. This dual capability allows HIA-NN to offer not just real-time metabolic data but also more accurate images, providing comprehensive visual and quantitative analyses that deliver deep insights into metabolic dynamics at both the molecular and visual levels.

Results, Conclusions, and Discussions: HIA-NN enhances metabolic research by integrating Stimulated Raman Scattering Imaging Technology (SRS) with advanced machine learning. This combination provides deeper insights into molecular dynamics related to aging and diseases. By targeting -CH and -CD vibrational regions, HIA-NN is able to efficiently identifies the organic compounds, contributing our understanding of metabolic disorders. We believe future enhancements in HIA-NN such as precision enhancement and broader spectral data application will expand its utility in bioimaging research and, and benefit more and more researchers.

Acknowledgements (Optional):