End-to-end lung nodule detection framework with model-based feature projection block

TL;DR

This paper introduces an end-to-end lung nodule detection framework that leverages a novel model-based feature projection block, achieving state-of-the-art results on the LUNA2016 dataset without needing false positive reduction steps.

Contribution

The paper presents a new nodule segmentation architecture with a model-based feature projection block integrated into a 3D-to-2D convolutional pipeline, simplifying the detection process.

Findings

Achieves 0.959 sensitivity on LUNA2016 dataset.

Eliminates the false positives reduction step.

Outperforms previous methods with state-of-the-art accuracy.

Abstract

This paper proposes novel end-to-end framework for detecting suspicious pulmonary nodules in chest CT scans. The method core idea is a new nodule segmentation architecture with a model-based feature projection block on three-dimensional convolutions. This block acts as a preliminary feature extractor for a two-dimensional U-Net-like convolutional network. Using the proposed approach along with an axial, coronal, and sagittal projection analysis makes it possible to abandon the widely used false positives reduction step. The proposed method achieves SOTA on LUNA2016 with 0.959 average sensitivity, and 0.936 sensitivity if the false-positive level per scan is 0.25. The paper describes the proposed approach and represents the experimental results on LUNA2016 as well as ablation studies.