One-class SVM for 3D region growing segmentation of liver in Computed Tomography

Improving automatic segmentation of liver tumor images using a deep learning model

“…Over the past few decades, the methods for liver segmentation in CT images were mainly designed based on the intensity, texture, and other information of abdominal CT images, including region growing, thresholding, and level set algorithms. Since the traditional region growing algorithm relies on artificially selected seed points, based on this, Thomaz et al optimized the traditional region growing algorithm [ 13 ], using the t position scale distribution to obtain the position and scale parameters of the target area for segmentation. This method does not rely on artificially selected seed points, but it does not have wide adaptability.…”