Encoder-decoder with dense dilated spatial pyramid pooling for prostate MR images segmentation

Authors

Lei Geng, Tianjin Polytechnic University, Tianjin Key Laboratory of Optoelectronic Detection Technology and Systems
Jia-Yan Wang, Tianjin Polytechnic University, Tianjin Key Laboratory of Optoelectronic Detection Technology and Systems
Zhitao Xiao, Tianjin Polytechnic University, Tianjin Key Laboratory of Optoelectronic Detection Technology and SystemsFollow
Jun Tong, University of Wollongong, Tianjin Key Laboratory of Optoelectronic Detection Technology and SystemsFollow
Fang Zhang, Tianjin Polytechnic University, Tianjin Key Laboratory of Optoelectronic Detection Technology and SystemsFollow
Jun Wu, Tianjin Polytechnic University, Tianjin Key Laboratory of Optoelectronic Detection Technology and SystemsFollow

RIS ID

138504

Publication Details

L. Geng, J. Wang, Z. Xiao, J. Tong, F. Zhang & J. Wu, "Encoder-decoder with dense dilated spatial pyramid pooling for prostate MR images segmentation," Computer Assisted Surgery, vol. 24, (Sup2) pp. 13-19, 2019.

Abstract

Automatic segmentation of prostate magnetic resonance (MR) images has great significance for the diagnosis and clinical application of prostate diseases. It faces enormous challenges because of the low contrast of the tissue boundary and the small effective area of the prostate MR images. In order to solve these problems, we propose a novel end-to-end professional network which consists of an Encoder-Decoder structure with dense dilated spatial pyramid pooling (DDSPP) for prostate segmentation based on deep learning. First, the DDSPP module is used to extract the multi-scale convolution features in the prostate MR images, and then the decoder is used to capture the clear boundary of prostate. Competitive results are produced over state of the art on 130 MR images which key metrics Dice similarity coefficient (DSC) and Hausdorff distance (HD) are 0.954 and 1.752 mm respectively. Experimental results show that our method has high accuracy and robustness.