BezierSeg: Parametric Shape Representation for Fast Object Segmentation in Medical Images

TL;DR

BezierSeg introduces a parametric shape model using Bezier curves for fast, accurate, and smooth lesion segmentation in medical images, addressing boundary artifacts common in pixel-wise methods.

Contribution

The paper presents BezierSeg, a novel contour-based segmentation method that ensures smooth, connected boundaries and allows easy manual refinement, with real-time performance.

Findings

Achieves competitive accuracy with pixel-wise models

Runs in real time on medical images

Provides smooth, continuous lesion boundaries

Abstract

Delineating the lesion area is an important task in image-based diagnosis. Pixel-wise classification is a popular approach to segmenting the region of interest. However, at fuzzy boundaries such methods usually result in glitches, discontinuity, or disconnection, inconsistent with the fact that lesions are solid and smooth. To overcome these undesirable artifacts, we propose the BezierSeg model which outputs bezier curves encompassing the region of interest. Directly modelling the contour with analytic equations ensures that the segmentation is connected, continuous, and the boundary is smooth. In addition, it offers sub-pixel accuracy. Without loss of accuracy, the bezier contour can be resampled and overlaid with images of any resolution. Moreover, a doctor can conveniently adjust the curve's control points to refine the result. Our experiments show that the proposed method runs in real time and achieves accuracy competitive with pixel-wise segmentation models.