Identification of Peritonitis Using Two-Stream Deep Spatial-Temporal Convolutional Networks

Abstract

Cine magnetic resonance imaging (MRI) analysis methods are used to evaluate intestinal peristalsis. However, the evaluation of intestinal peristalsis by MRI is subjective, time-consuming, and not reproducible, which are recognized as an important issue that needs to be addressed. In our previous work, we used a deep optical flow network (DOFN) to extract temporal-spatial features of intestinal movements and differentiate peritonitis from intestinal peristalsis. However, since the DOFN is based on the image difference of two neighboring frames, it lacks texture and spatial information of small bowels. To solve these problems, this paper proposed a new model with two-stream deep spatial-temporal convolutional networks (two-stream DSTCN) consisting of optical flow stream (i.e, DOFN) and dynamic image stream. The proposed method is an improved version of our DOFN by introducing a dynamic image stream to extract temporal-spatial features from cine MR images. The final result is obtained by the average fusion of the two streams. The accuracy is improved by about 3% with the proposed method.