Intelligent Recognition of Antigen Detection Reagent for Corona virus based on Improved Text Recognition

In the wake of the post-epidemic era, the gradual standardization of coronavirus detection has given rise to concerns about the efficiency of corona virus self-test result identification. This study primarily employs a scene text recognition technology based on DBNet and SVTR to discern the outcomes of corona virus detection. The methodology involves locating and recognizing the C-line and T-line within images of corona virus antigen detection reagents. By rectifying image positioning, the text region is initially identified and screened using a DBNet-based text recognition algorithm. Subsequently, the SVTR model is employed to pinpoint key text elements. The detection area is assessed through text-based positioning, followed by color extraction and straight line detection to ascertain the ultimate result. Experimental findings underscore the high recognition efficiency and accuracy of the DBNet and SVTR models in this context, suggesting their potential for widespread adoption.