Gradient-adaptive Nonlinear Sharpening for Dental Radiographs

Unsharp Masking is a popular image processing technique used for improving the sharpness of structures on dental radiographs. However, it produces overshoot artefact and intolerably amplifies noise. On radiographs, often resembles indications prosthesis misfit, pathosis, pathological features associated with restorations. A noise- robust alternative to algorithm, termed Gradient-adaptive Nonlinear Sharpening (GNS) which free from discontinuity artefacts, proposed in this paper. In GNS, product arbitrary scalar as ‘scale’ difference between output Adaptive Edge Smoothing Filter (AESF) input image, weighted by normalized gradient magnitude added image. AESF locally-adaptive 2D Gaussian smoothing kernel whose variance directly proportional local value magnitude. The dataset employed paper downloaded Mendeley data repository having annotated panoramic radiographs 116 patients. values Saturation Evaluation Index (SEI), Sharpness Ridges (SOR), Model Based Contrast Metric (EMBCM), Visual Information Fidelity (VIF) exhibited are 0.0048 ± 0.0021, 4.4 × 1013 3.8 1013, 0.2634 0.2732 0.9898 0.0122. these quality metrics corresponding GNS 0.0042 0.0017, 2.2 1.8 0.5224 0.1825, 1.0094 0.0094. lower SEI SOR higher EMBCM VIF, compared Masking. Lower SOR, respectively indicate that saturation edges images less affected Higher confirm haloes thin sharp sharpened good information fidelity.