Automatic Range Adjustment of the Fluorescence Immunochromatographic Assay Based on Image Processing

Ruixin Jiang, Huihuang Wu, Jianpeng Yang, Haiyan Jiang, Min Du, Mang I Vai, Siohang Pun, Yueming GaoSensors 2019 · DOI

TL;DR

This paper introduces an automatic exposure time adjustment mechanism for fluorescence immunochromatographic assay (FICA) readers that use common CMOS image sensors. By dynamically adjusting exposure, the system significantly expands the detection range and improves sensitivity without requiring expensive high-precision sensors, making point-of-care diagnostics more accessible.

What Problem Does This Paper Address?

Fluorescence immunochromatographic assays (FICAs) are powerful tools for rapid, sensitive detection in medical, food safety, and environmental monitoring. While image processing-based FICA readers offer portability and ease of use, their detection range is often limited by the inherent gray level limitations of standard CMOS image sensors. This means that very low or very high concentrations of analytes might fall outside the measurable range, leading to inaccurate or unavailable results. High-precision sensors that could overcome this limitation are prohibitively expensive, hindering the widespread adoption of these diagnostic tools. The authors aim to bridge this gap by enhancing the performance of affordable FICA readers.

Methods

The core of this work lies in developing an automatic range adjustment mechanism. The authors leverage the linear relationship between image gray value and exposure time for a given excitation light intensity.

  1. Exposure Time Adjustment: They identified that the gray value (G) of the image captured by the CMOS sensor is linearly proportional to the exposure time (t) within a certain range (G = 0.547t + 3.480, R² = 0.999). This linear relationship is crucial.
  2. Range Control Function: A range control function was established to adjust the exposure time. The goal is to drive the gray mean value of the target area (e.g., the test line on the strip) to a maximum allowed value (GM). The new exposure time (t') is calculated as t' = Kt, where K is the ratio of the maximum gray value to the current target gray value (GM/G). This ensures the target gray value is within the detectable range.
  3. Image Processing Pipeline: The FICA reader employs a standard image processing pipeline: image acquisition, denoising (using mathematical morphology for smoothing and edge refinement), and segmentation (using an improved Otsu method based on image entropy for robust thresholding, which is less sensitive to illumination and contrast changes).

Datasets

Not discussed in the paper. The experiments were conducted using prepared C-reactive protein (CRP) solutions at various concentrations and FICA strips.

Key Insights and Ideas

  • Exposure Time as a Lever for Range Expansion: The most significant insight is that by intelligently controlling the camera's exposure time, the dynamic range of a FICA reader can be significantly expanded. This is a clever way to overcome the limitations of sensor bit depth without resorting to more expensive hardware.
  • Linearity is Key: The authors demonstrate a strong linear relationship between image gray value and exposure time. This linearity is fundamental to their proposed automatic adjustment method, allowing for predictable and controllable adjustments.
  • No Impact on T/C Ratio: Crucially, they mathematically show that adjusting exposure time does not alter the final detection result (the T/C ratio), which is the critical metric for quantitative analysis in FICA. This means the adjustment enhances the ability to measure, not the measurement itself.
  • Photobleaching Mitigation: By fixing the excitation light intensity at a low, constant value and adjusting exposure time, the authors mitigate the risk of photobleaching, a known issue with high-intensity illumination that can degrade fluorescent signals.

Key Results

  • Detection Range Expansion: The maximum detection range was expanded from 1.95–125 µg/mL to 0.98–512 µg/mL for CRP detection. This represents a significant increase in the upper detection limit.
  • Sensitivity Improvement: The detection sensitivity showed a onefold increase, and the upper detection limit showed a twofold increase after the method was implemented.
  • High Linearity and Accuracy: The fitting degree (R²) for both linearity and accuracy experiments reached 0.999, indicating excellent performance and strong correlation with a reference instrument (ESEQuant LFR).
  • Consistency: The adjusted FICA reader showed good consistency with the ESEQuant LFR, a commercial reader.

Limitations and Future Work

  • Limitations: The paper acknowledges that the gray level of common 8-bit image sensors inherently limits the detection range. While their method expands this range, there will still be an ultimate limit. The current system relies on a PC for processing, which limits portability.
  • Future Work: The authors propose developing a detection system on a mobile platform. This would eliminate the PC dependence, reduce the overall volume, and further enhance the portability of the FICA reader.

Summary updated

Abstract

As an emerging technology, fluorescence immunochromatographic assay (FICA) has the advantages of high sensitivity, strong stability and specificity, which is widely used in the fields of medical testing, food safety and environmental monitoring. The FICA reader based on image processing meets the needs of point-of-care testing because of its simple operation, portability and fast detection speed. However, the image gray level of common image sensors limits the detection range of the FICA reader, and high-precision image sensors are expensive, which is not conducive to the popularization of the instrument. In this paper, FICA strips' image was collected using a common complementary metal oxide semiconductor (CMOS) image sensor and a range adjustment mechanism was established to automatically adjust the exposure time of the CMOS image sensor to achieve the effect of range expansion. The detection sensitivity showed a onefold increase, and the upper detection limit showed a twofold increase after the proposed method was implemented. In addition, in the experiments of linearity and accuracy, the fitting degree (<i>R</i><sup>2</sup>) of the fitted curves both reached 0.999. Therefore, the automatic range adjustment method can obviously improve the detection range of the FICA reader based on image processing.
BibTeX
@inproceedings{jiang2019automatic,
  author = {Ruixin Jiang and Huihuang Wu and Jianpeng Yang and Haiyan Jiang and Min Du and Mang I Vai and Siohang Pun and Yueming Gao},
  title = {{Automatic Range Adjustment of the Fluorescence Immunochromatographic Assay Based on Image Processing}},
  booktitle = {Sensors},
  year = {2019},
  doi = {10.3390/s20010209},
}
Figure