A fast two-stage active contour model for intensity inhomogeneous image segmentation.

Area of Science:

• Computer Vision
• Image Processing
• Machine Learning

Background:

• Intensity inhomogeneity is a common challenge in image segmentation.
• Existing methods often struggle with accuracy and computational efficiency for such images.

Purpose of the Study:

• To develop a fast and accurate two-stage image segmentation method for intensity inhomogeneous images.
• To improve upon existing active contour models and clustering techniques.

Main Methods:

• A two-stage segmentation approach utilizing a local region-based active contour model with an exponential family energy function.
• Stage 1: Preliminary segmentation of down-sampled images using local correntropy-based K-means clustering.
• Stage 2: Precise segmentation of original images using an improved local correntropy-based K-means model, initialized by the Stage 1 result.
• Energy function convergence using Riemannian steepest descent method for global minima attainment with fewer iterations.

Main Results:

• The proposed method achieves accurate segmentation of intensity inhomogeneous images.
• Demonstrates higher efficiency and lower computational complexity compared to state-of-the-art methods.
• Experimental results on synthetic and real images validate the method's effectiveness.

Conclusions:

• The presented two-stage method offers a significant advancement in segmenting intensity inhomogeneous images.
• It provides a robust and efficient solution suitable for various image processing applications.