Interpretation of urine dipstick results based on color similarity using linear interpolation curve fitting

Abstract

Automation of urine dipstick reagent results interpretation is influenced by color space selection and the method to determine color similarity between a tested color and the target color chart. Color similarity among colors can be evaluated by calculating their Euclidean distance in a color space; however it only calculates the distances between a tested color and a series of target color in a color chart, producing a stepwise quantification. It is not suitable if a value between two target colors should be estimated, producing a smoother quantification. This research enhances the method for urine dipstick color interpretation. A stepwise linear interpolation algorithm in CIELAB color space is used. It produces smoother quantification and a possibility to verify whether the tested reagent is belong to the certain color chart or not. A formula resulted from our algorithm can be easily implemented for interpreting the score of the tested reagent. The result of this automation method is evaluated by comparing the result with a visual interpretation and color interpretation by using Euclidean distance only. The result shows a promising method in an automation process for dipstick urinalysis.