Related Experiment Video
Updated: Jul 13, 2025
06:12
Multimodal Analytical Platform on a Multiplexed Surface Plasmon Resonance Imaging Chip for the Analysis of Extracellular Vesicle Subsets
Published on: March 17, 2023
1.5K
Deep Learning Approach for the Localization and Analysis of Surface Plasmon Scattering
Jongha Lee1, Gwiyeong Moon1, Sukhyeon Ka1
1School of Electrical and Electronic Engineering, Yonsei University, Seoul 03722, Republic of Korea.
Sensors (Basel, Switzerland)
|October 14, 2023
Summary
Deep learning, using the Y-Net model, enhances Surface Plasmon Resonance Microscopy (SPRM) for label-free imaging. This method accurately detects and analyzes scatterers in one shot, improving SPRM
Area of Science:
- Optics and Photonics
- Biomedical Imaging
- Machine Learning
Background:
- Surface Plasmon Resonance Microscopy (SPRM) offers label-free imaging capabilities by integrating traditional microscopy with surface plasmon properties.
- Current SPRM methods face challenges in efficient detection and analysis of scattering parameters, particularly in noisy environments.
- There is a need for advanced analytical techniques to enhance the resolution and speed of SPRM imaging.
Purpose of the Study:
- To introduce a novel deep learning-based approach utilizing the Y-Net convolutional neural network for improved SPRM image analysis.
- To develop a one-shot image analysis method for estimating scattering parameters, including scatterer location, from SPRM data.
- To validate the efficacy of the deep learning method in enhancing SPRM's detection and characterization capabilities.
Main Methods:
- Implementation of the Y-Net convolutional neural network model for processing SPRM images.
- Development of a machine learning-based image analysis technique for one-shot estimation of scattering parameters.
- Assessment of the method by applying it to SPRM images and comparing reconstructed images with original data.
Main Results:
- The deep learning approach demonstrated high accuracy in localizing scatterers and predicting scattering object variables, even in noisy conditions.
- The Y-Net model successfully reconstructed SPRM images from network outputs, showing good agreement with original images.
- The method proved effective in achieving one-shot localization and characterization of scatterers, significantly boosting SPRM's detection capabilities.
Conclusions:
- Deep learning, specifically the Y-Net model, offers a powerful tool for enhancing Surface Plasmon Resonance Microscopy.
- The developed one-shot analysis method substantially improves the accuracy and efficiency of scatterer detection and characterization in SPRM.
- This approach holds significant potential for advancing label-free imaging techniques in various scientific and biomedical applications.
