Abstract
The automatic detection and segmentation of liver lesion is useful in many clinical application, whereas it remains a challenging task due to the largely varied shape, size and texture of the diseased masses. In this paper, we present a cascade learning approach comprising multiple classifiers for the detection of two different types of solid liver lesions, hypodense and hyperdense lesions. In particular, we propose an efficient gradient based locally adaptive segmentation method for the solid lesions, where the segmentation results are used to extract shape features to boost up the detection performance. The proposed method is validated on a total of 660 volumes with 1,302 hypodense lesions, and 234 volumes with 328 hyperdense lesions. The experimental results show a resulting 90% detection rate at 1.01 false positives per volume for hypodense lesion and 1.58 false positives per volume for hyperdense lesion, respectively, using three fold cross validation.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call