Accuracy of Ray-Ban Meta Smart Glasses, ChatGPT, and Newly Graduated Dentists: A Game Changer in Dental Material Identification (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Wearable technology and artificial intelligence (AI) are reshaping the healthcare and education sectors. This study compared the abilities of Ray-Ban Meta smart glasses, ChatGPT, and newly graduated dentists in identifying dental materials from images. </sec> <sec> <title>OBJECTIVE</title> The primary aim was to evaluate which group achieved superior accuracy in recognizing items such as resin cements, bonding agents, and composite tubes and to understand their respective uses. </sec> <sec> <title>METHODS</title> This was a randomized controlled trial that involved 60 participants divided into three equal groups. Group A utilized Ray-Ban Meta smart glasses equipped with AR-enhanced material identification features; Group B relied on ChatGPT via a tablet interface for AI-generated identification and descriptions; and Group C consisted of newly graduated dentists who utilized their foundational knowledge and practical experience. High-resolution images of dental material boxes, including resin cements, bonding agents, composite tubes, etching gels, dental liners and bases, and polishing kits were used as test items. These images simulate real-world identification scenarios (Figure 1). Participants were shown the images sequentially and were asked to identify the material's name, describe its primary use, and provide additional relevant information about its application or limitations. The identification process varied by group, where Group A benefited from AR overlays and contextual feedback, Group B received text-based material identification from ChatGPT, and Group C relied solely on pre-existing knowledge. Evaluation criteria The following evaluation criteria were used to ensure objective and consistent assessment across all groups: • Accuracy: Each correctly identified material name and primary use was awarded one point, each. A maximum score of two points per material was possible. • Efficiency: The time taken to identify each material was recorded in seconds. • Usability: Survey scores assessed ease of use, confidence, and satisfaction on a 5-point Likert scale. Usability survey questions Participants completed a usability survey after the task, and the questions included: 1. How easy was it to understand and use the tool for identifying materials? 2. How confident did you feel in your answers during the task? 3. How satisfied were you with the assistance provided by the tool or method? 4. Did you encounter any difficulties or challenges during the identification process? 5. Would you recommend this tool or method for use in clinical or educational settings? </sec> <sec> <title>RESULTS</title> The results revealed that the Ray-Ban Meta smart glasses achieved the highest accuracy and efficiency among the groups. The mean accuracy for Group A was 93.5%, with a standard deviation of 3.2%. Group B showed an accuracy of 87.1% with a standard deviation of 5.4%, whereas Group C had the lowest accuracy, at 79.6% with a standard deviation of 6.7%. In terms of efficiency, Group A identified materials with a mean time of 12.4 seconds per item and standard deviation of 1.8 seconds. Group B required a mean time of 15.2 seconds with a standard deviation of 2.5 seconds, and Group C took the longest, with a mean time of 18.9 seconds and a standard deviation of 3.3 seconds. The usability feedback further highlighted the advantages of Ray-Ban Meta smart glasses. Group A had the highest ease of use (4.8), confidence (4.5), and satisfaction (4.6) scores. Group B achieved moderate scores, with ease of use at 4.2, confidence at 3.9, and satisfaction at 4.1. Group C scored the lowest across all metrics, with ease of use at 3.7, confidence at 3.5, and satisfaction at 3.6. </sec> <sec> <title>CONCLUSIONS</title> By analyzing accuracy, efficiency, and user experience, this study provides insights into the potential of augmented reality and AI technologies in enhancing dental education and practice. </sec>

Ähnliche Arbeiten