Predicting Final Restoration Color Using Neural Network Models: The Impact of Substrate Lightness Versus Ceramic...

Area of Science:

• Dental Materials Science
• Color Science in Dentistry
• Artificial Intelligence in Prosthodontics

Background:

• Accurate color matching is crucial for esthetic dental restorations.
• Lithium disilicate ceramics offer excellent mechanical and esthetic properties.
• Predicting color outcomes requires understanding complex interactions between material and environmental factors.

Purpose of the Study:

• To evaluate the influence of background color, ceramic shade, translucency, and thickness on lithium disilicate color matching.
• To develop and utilize a neural network model for predicting optimal shade matching parameters.
• To identify key factors influencing color discrepancies in ceramic restorations.

Main Methods:

• A neural network model was employed to analyze the effects of ceramic shade, translucency, and thickness.
• Various background colors, including enamel and dentin shades, were tested.
• Color measurements (L*, a*, b*) were recorded using a spectrophotometer (CM-700d, Konica Minolta).

Main Results:

• Background color significantly influenced color matching (p < 0.001).
• The neural network model explained 71.45% of the variance in final color values.
• Initial L* value, ceramic thickness, and translucency were identified as key predictors of color matching.

Conclusions:

• Background color and ceramic thickness are critical factors for minimizing color discrepancies.
• The developed neural network model shows potential for predicting ceramic restoration color outcomes.
• Neural networks can assist clinicians in material selection and parameter optimization for esthetic results.