3D Context Enhanced Region-based Convolutional Neural Network for End-to-End Lesion Detection

TL;DR

This paper introduces 3DCE, a novel end-to-end CNN framework that effectively incorporates 3D context for lesion detection in CT scans, improving accuracy over traditional 2D methods.

Contribution

The paper presents a new 3D context enhanced CNN architecture for lesion detection that is easy to train and applicable to multiple lesion types in a single model.

Findings

3DCE outperforms existing methods on the DeepLesion dataset.

The framework is end-to-end trainable and efficient.

Code is publicly available for reproducibility.

Abstract

Detecting lesions from computed tomography (CT) scans is an important but difficult problem because non-lesions and true lesions can appear similar. 3D context is known to be helpful in this differentiation task. However, existing end-to-end detection frameworks of convolutional neural networks (CNNs) are mostly designed for 2D images. In this paper, we propose 3D context enhanced region-based CNN (3DCE) to incorporate 3D context information efficiently by aggregating feature maps of 2D images. 3DCE is easy to train and end-to-end in training and inference. A universal lesion detector is developed to detect all kinds of lesions in one algorithm using the DeepLesion dataset. Experimental results on this challenging task prove the effectiveness of 3DCE. We have released the code of 3DCE in https://github.com/rsummers11/CADLab/tree/master/lesion_detector_3DCE.