Abstract
This paper describes a method for detecting small nodules from three-dimensional chest X-ray CT images. The proposed method consists of two submodules: (a) initial detection of nodule candidates using shape features of lung nodules and (b) reduction process of false positive regions using a minimum directional difference filter called Min-DD from the initial candidates. The performance of the proposed method was evaluated by using six cases of chest X-ray CT images including five abnormal cases where multiple lung cancers are observed. The experimental results showed that sensitivity and FP regions are 71% and 7.4 regions in average per case for nodules of 2 mm in diameter (361 regions in total).
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have