Optimization of OpenCV based spot identification method for surface plasmon resonance imaging

Zhiyou Wang,Feiyu Liu,Wenxuan Xiao,Zhewen Fang,Chang Ou

DOI: https://doi.org/10.1063/5.0192315

IF: 1.697

2024-02-01

AIP Advances

Abstract:In this work, we focus on the OpenCV based microarray recognition method for Surface Plasmon Resonance Imaging (SPRi), proposing the hit-ratio of global light pixels and coverage of the potential spots in a microarray as the criteria for identification evaluation in SPRi data. We optimized the design of the ellipse fitting strategy by analyzing the impact of different parameters in the method. After optimization of the parameters, the accuracy of microarray recognition was successfully increased to over 90%. This work not only contributes to reducing errors in microarray signal extraction and improving signal processing quality but also has significant implications for applying computer graphic technology in high-throughput biochemical analysis.

materials science, multidisciplinary,physics, applied,nanoscience & nanotechnology

What problem does this paper attempt to address?

The paper primarily addresses the issue of microarray recognition methods in Surface Plasmon Resonance Imaging (SPRi) technology. Specifically, the research team proposed an optimized OpenCV-based method for spot targets in SPRi data (usually referring to detection points of biomolecular interactions) and improved the design of the ellipse fitting strategy by adjusting key parameters to enhance the accuracy of microarray recognition.

The core contributions of the paper are:
1. **Proposed Evaluation Criteria**: Using the hit rate of global highlight pixels and the coverage of potential points in the microarray as the standards for recognition evaluation.
2. **Parameter Optimization**: By analyzing the impact of different parameters, the steps of image enhancement, edge continuity compensation, edge sharpening, binarization, and contour detection and fitting were optimized.
3. **Performance Improvement**: After parameter optimization, the accuracy of microarray recognition successfully increased to over 90%, significantly reducing errors in the signal extraction process and improving the quality of signal processing.

In addition, the study compared the recognition effects under different parameter combinations and ranked the parameter combinations based on hit rate and coverage, identifying the optimal parameter combination scheme. These optimizations not only help reduce errors in microarray signal extraction but also have significant implications for the application of computer graphics technology in high-throughput biochemical analysis.