Edge extraction method for medical images based on improved local binary pattern combined with edge-aware filtering

Edge-based processing and analysis of medical images are indispensable in modern diagnosis and the application value of edge extraction technology is rising with this tide. For medical images containing redundant noise, blurred details, and low contrast, a robust edge extraction method based on edge-aware filtering and improved local binary pattern (EF-ALBP) is proposed in this paper. EF-ALBP contains two parts: the edge-aware filtering (EF) is proposed to suppress noise and enhance contrast while preserve edges, and ALBP is used to extract the crucial edge features of the previous step results accurately by introducing an accumulation function into local binary pattern. Quantitative analyses and visual evaluation for experimental results on X-ray, CT, and MRI images from medical image datasets demonstrate that the proposed method is competitive in robustness and outperforms those of the popular methods.