An intelligent signal processing strategy for accurate quantification of composite signals in background fluorescence-quenching lateral flow immunoassay for folic acid.

Background fluorescence-quenching lateral flow immunoassay (BF-LFIA) holds great potential for high-sensitivity detection owing to its unique “turn-on” mechanism. However, its characteristic composite signals, comprising “quenched dark zones” and “fluorescent bright zones,” exhibit high heterogeneity and dynamic evolution, severely restricting the accuracy and universality of automated quantification. To overcome this challenge, a portable analysis system was constructed and a dedicated intelligent image analysis framework is proposed. This framework integrates three core algorithm modules: (1) Synergistic Color-Space Segmentation (SCSS), which leverages the complementary advantages of HSV and CIELAB spaces to achieve robust extraction of fluorescent signal regions; (2) Adaptive Spatial Localization (ASL), which employs pixel density projection and spatial constraint mechanisms to precisely lock onto the regions of interest (ROI) for Test (T) and Control (C) lines; and (3) Adaptive Fusion Quantification (AFQ), which integrates multi-dimensional color features via a dynamic weight allocation mechanism. Validated by a Random Forest model with 5-fold cross-validation, the AFQ algorithm demonstrated robustness in the precise resolution of composite signals. Using folic acid (FA) as the analyte, the system achieved ultrasensitive quantitative detection within the 0–300 ng/mL range. The concentration-response curve, fitted by a four-parameter logistic model, exhibited excellent linearity (R2 = 0.9996). The proposed strategy effectively resolves the quantification challenge of composite signals in BF-LFIA platforms, providing a powerful tool for high-performance point-of-care testing (POCT) and offering a general methodological reference for other biomarker detections based on background fluorescence-quenching mechanisms.