Liver Lesion Detection from Weakly-labeled Multi-phase CT Volumes with a Grouped Single Shot MultiBox Detector

TL;DR

This paper introduces a modified Single Shot MultiBox Detector with grouped convolutions for liver lesion detection in multi-phase CT scans, achieving high accuracy and fast inference, suitable for real-world clinical use.

Contribution

The study presents a novel grouped convolution approach within SSD for multi-phase CT data, improving detection accuracy and efficiency over existing models.

Findings

Achieved 53.3% average precision in liver lesion detection.

Model runs in under three seconds per volume, enabling near real-time analysis.

Outperforms original SSD and recent variants on a clinical CT dataset.

Abstract

We present a focal liver lesion detection model leveraged by custom-designed multi-phase computed tomography (CT) volumes, which reflects real-world clinical lesion detection practice using a Single Shot MultiBox Detector (SSD). We show that grouped convolutions effectively harness richer information of the multi-phase data for the object detection model, while a naive application of SSD suffers from a generalization gap. We trained and evaluated the modified SSD model and recently proposed variants with our CT dataset of 64 subjects by five-fold cross validation. Our model achieved a 53.3% average precision score and ran in under three seconds per volume, outperforming the original model and state-of-the-art variants. Results show that the one-stage object detection model is a practical solution, which runs in near real-time and can learn an unbiased feature representation from a large-volume real-world detection dataset, which requires less tedious and time consuming construction of the weak phase-level bounding box labels.