Dual-functionalized bilayer colorimetric hydrogel label combined with deep learning for intelligent monitoring of chicken freshness.

A dual-functional bilayer colorimetric hydrogel label was engineered in this study. It consists of a ZnO/SA-AG UV-shielding upper layer and a liposome-loaded natural pigment indicator layer with curcumin, anthocyanin and betanin, which are assembled into a 3 × 1 window array to construct three independent chromogenic units. The optimized 0.1 wt% ZnO provided strong UV blocking while maintaining transparency. Liposomal encapsulation enhanced pigment stability, yielding encapsulation efficiencies of 81.91 ± 2.10% for curcumin, 69.12 ± 1.90% for anthocyanin, and 71.08 ± 1.70% for betanin. The bilayer liposome configuration exhibited superior ammonia responsiveness (S = 37.77%, 33.31%, 75.50%) and strong photostability, with color-difference increases of only 10.22%, 9.94%, 14.91% after 10-day simulated daylight exposure. During refrigerated chicken storage, the total color difference correlated strongly with TVB-N (R2 = 0.9173), enabling accurate freshness classification. Coupled with an optimized EfficientNetV2-S model, the system achieved 0.972 accuracy and Macro-F1 for a three-class classification task (fresh/less-fresh/spoiled). This work establishes a compact, sensitive, and intelligent three-pigment, three-zone material-algorithm platform suitable for practical smart-packaging applications.