Abstract
Introduction: Cavities that are difficult to detect through visual inspection pose risks to dental health. To address this, we propose a detection method using the Split Spectrum Optical Attenuation Coefficient (OAC), capitalizing on the wavelength-dependent properties of carious dental structures. Methods: The complete OCT spectrum is divided into four spectral segments, each independently reconstructed and analyzed for OAC. OAC values are then compared across segments to differentiate between healthy and carious tissue. Experiments were conducted on ex vivo human teeth, with carious areas marked by dental professionals. The accuracy of the proposed method was evaluated and compared with the full-spectrum OAC method. Results: For healthy enamel, OAC values ranged from 0.2 to 0.4 mm⁻¹, remaining consistent across spectral bands. In carious enamel, OAC increased significantly (0.5-2.5 mm⁻¹), with a 60% rise in short-wave versus long-wave OAC. Healthy dentin showed OAC values from 0.3 to 0.7 mm⁻¹, whereas carious dentin reached 0.8-1.2 mm⁻¹, with a 50% increase in short-wave OAC compared to long-wave. The average false positive rate of the method proposed in this paper is 0.9%, significantly lower than the 1.8% of the traditional OAC method (t-test, n=16, p=0.013). The average false negative rates of both methods are around 1%, with no significant difference. Conclusion: Findings indicate that the split-spectrum OAC method can effectively identify caries, with higher accuracy and specificity compared to the traditional OAC method.