A Bottom-Up Approach for Pancreas Segmentation Using Cascaded Superpixels and (Deep) Image Patch Labeling.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Radiology

Background:

• Robust organ segmentation is crucial for medical applications like computer-aided diagnosis and surgical assistance.
• Accurate segmentation of organs with high anatomical variability, such as the pancreas, remains challenging.

Purpose of the Study:

• To develop an automated, bottom-up approach for accurate pancreas segmentation in abdominal computed tomography (CT) scans.
• To improve computational efficiency and numerical stability compared to existing segmentation methods.

Main Methods:

• The approach involves decomposing CT slices into superpixels, computing pancreas class probability maps using dense patch labeling (Random Forest and Deep Convolutional Neural Networks), and classifying superpixels.
• A four-step process includes superpixel decomposition, probability map computation, superpixel classification using pooled features, and connectivity-based post-processing.

Main Results:

• The method achieved a Dice coefficient of 70.7% and a Jaccard index of 57.9%, outperforming other state-of-the-art methods.
• Computational time was significantly reduced to 6-8 minutes per case, compared to over 10 hours for other methods.
• The deep patch labeling approach demonstrated greater numerical stability with smaller performance metric standard deviations.

Conclusions:

• The proposed automated bottom-up pancreas segmentation method is highly accurate and computationally efficient.
• It significantly outperforms traditional multi-atlas label fusion (MALF) approaches for pancreas segmentation.
• This technique offers a promising advancement for computer-aided diagnosis and quantitative analysis in medical imaging.