Evaluation of Color Matching Accuracy Using Artificial Intelligence Applications and a Spectrophotometer: A Photometric Analysis

Abstract

Statement of problem. The use of artificial intelligence (AI) applications in dentistry is increasingly widespread. However, studies evaluating their success in tooth color matching are limited. Purpose. The purpose of this in vitro study was to compare the color matching accuracy of AI applications and a spectrophotometer with photometric analysis. Material and methods. To evaluate the success of color matching, 13 acrylic resin teeth were selected from the VITA Toothguide 3D-Master color scale (n=13). Test groups included ChatGPT-4 (Group C), Gemini 1.5 Pro (Group G), and Easyshade Spectrophotometer (Group ES). Photographs of the test samples were made with a mobile phone camera under standard conditions and uploaded to both AI applications for color matching. Color measurements were made with the ES, and the data were recorded. Additionally, all teeth were photographed using a digital single lens reflex camera with a polarizing filter and gray card. After calibration, L*, a*, and b* values were extracted from the middle third using an image analysis Lightroom program. The CIEDE2000 (Delta E-0(0)) formula was used to calculate color differences between reference and test group values. Data distribution was assessed with the Shapiro-Wilk test, and group comparisons were performed using 1-way ANOVA (alpha=.05). Results. The mean color difference values of Groups C, G, and ES were 2.84, 1.94, and 0.7, respectively. While Group C and Group G were statistically similar (P=.844), significant differences were found between Group ES and both Group C (P=.017) and Group G (P=.029) Conclusions. The Easyshade spectrophotometer was found to be more successful than AI applications in color matching. AI color matching has demonstrated success above the clinically acceptable threshold (Delta E-0(0)>1.8).