Region-Manipulated Fusion Networks for Pancreatitis Recognition

TL;DR

This paper introduces a novel Region-Manipulated Fusion Network (RMFN) for automatic pancreatitis recognition in CT images, effectively highlighting lesion regions despite their subtle and variable appearance.

Contribution

The paper proposes a flexible region-manipulated scheme integrated into existing neural networks to improve local lesion detection for pancreatitis recognition.

Findings

Effective lesion highlighting in CT images demonstrated

Improved recognition accuracy on real hospital data

Method adaptable to networks like AlexNet and VGG

Abstract

This work first attempts to automatically recognize pancreatitis on CT scan images. However, different form the traditional object recognition, such pancreatitis recognition is challenging due to the fine-grained and non-rigid appearance variability of the local diseased regions. To this end, we propose a customized Region-Manipulated Fusion Networks (RMFN) to capture the key characteristics of local lesion for pancreatitis recognition. Specifically, to effectively highlight the imperceptible lesion regions, a novel region-manipulated scheme in RMFN is proposed to force the lesion regions while weaken the non-lesion regions by ceaselessly aggregating the multi-scale local information onto feature maps. The proposed scheme can be flexibly equipped into the existing neural networks, such as AlexNet and VGG. To evaluate the performance of the propose method, a real CT image database about pancreatitis is collected from hospitals \footnote{The database is available later}. And experimental results on such database well demonstrate the effectiveness of the proposed method for pancreatitis recognition.