Pulmonary Nodule Detection in CT Images using Dual Path U-Net and Multiscale Region Proposal Network

xu jiahua, Philipp Ernst, Eric Liu, Andreas Nürnberger

Keywords: Pulmonary nodule detection, Dual Path U-Net, Region Proposal Network

TL;DR: This paper proposes a novel deep learning model based on a Dual Path network in a U-Net structure generating multiscale feature maps as well as taking advantage of having 2.5D input to provide better contextual information.

Abstract: Pulmonary cancer is one of the most commonly diagnosed and deadly cancers and often diagnosed by incidental findings with computed tomography. Automated pulmonary nodule detection is an essential part of the computer-aided diagnosis, which is still facing great challenges and difficulties to quickly and accurately locate the exact nodules' positions. This paper proposes a novel deep learning model based on a Dual Path network in a U-Net structure generating multiscale feature maps as well as taking advantage of having 2.5D input to provide better contextual information. An extended upsampling strategy is proposed to minimize the ratio of false positives and maximize the sensitivity for lesion detection of nodules. The results show that our new upsampling strategy improves the performance by having 85.3% sensitivity at 4 FROC per image compared to 84.2% for the regular upsampling strategy as well as 81.2% for VGG16 based Faster-R-CNN.

Registration: I acknowledge that publication of this at MIDL and in the proceedings requires at least one of the authors to register and present the work during the conference.

Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.

Paper Type: both

Primary Subject Area: Detection and Diagnosis

Secondary Subject Area: Segmentation