A Bayesian Approach for Liver Analysis: Algorithm and Validation Study

Abstract

Abstract. We present a new method for the simultaneous, nearly automatic segmentation of liver contours, vessels, and metastatic lesions from abdominal CTA scans. The method repeatedly applies multi-resolution, multi-class smoothed Bayesian classification followed by morphological adjustment and active contours refinement. It uses multi-class and voxel neighborhood information to compute an accurate intensity distribution function for each class. The method requires only one or two user-defined voxel seeds, with no manual adjustment of internal parameters. A retrospective study on two validated clinical datasets totaling 56 CTAs was performed. We obtained correlations of 0.98 and 0.99 with a manual ground truth liver volume estimation for the first and second databases, and a total score of 67.87 for the second database. These results suggest that our method is accurate, efficient, and robust to seed selection compared to manually generated ground truth segmentation and to other semi-automatic segmentation methods.

Automated segmentation and quantification of liver and spleen from CT images using normalized probabilistic atlases and enhancement estimation

“…User interaction was requested for the iterative Bayesian approach proposed in Ref. 27. A fast hierarchical model using marginal space learning was introduced in Ref.…”

Liver tumour segmentation using contrast-enhanced multi-detector CT data: performance benchmarking of three semiautomated methods

“…This fully automatic method was reported to achieve a volume overlap of 94.2% and an accuracy of 3.7 mm for liver surface segmentation, and the sensitivity and specificity for tumour lesion detection were 82.6% and 87.5%, respectively [24]. In another approach, researchers used multiple class-based Bayesian classification [25] and graph-cut [26] for the simultaneous segmentation of liver parenchyma, lesions and vessels. According to our experience, this method may only be applicable to images from the portal venous phase with homogeneous hypodense liver lesions.…”

Liver tumors segmentation from CTA images using voxels classification and affinity constraint propagation

“…The automatic methods did not require any user interaction but had limited accuracy. We presented a nearly automatic method in which only a couple of seeds are required to initialize the segmentation algorithm [17,18]. This user interaction is appropriate for routine clinical use but yielded results with limited accuracy.…”