Share Email Print
cover

Proceedings Paper • new

An efficient deep neural network for classification in esophageal cancer histopathological images (CNN+LSTM)
Author(s): Jing Ge; Ying Xie; Di Wang
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Esophageal cancer is one of the leading causes of mortality worldwide. Early diagnostics are imperative of improving the chances of correct treatment and survival. Pathology examination requires time consuming scanning while often leads to a disagreement among pathologists. The computer-aided diagnosis systems are critical to improve the diagnostic efficiency and reduce the subjectivity and error of human. In this paper, a deep learning–based approach is proposed to classify the H&E stained histopathological images of esophageal cancer. The histopathological images are firstly normalized to correct the color variations during slide preparation. Then one image patch is cut into five pieces and gets five corresponded features via convolutional neural networks (CNNs). Subsequently, the five features of one image patch are fed into the Long Short-Term Memory (LSTM) model for further feature extraction and integrated as one for the last classification. The experiment results have demonstrated the proposed approach is effective to classify the esophageal histopathological images with the accuracy of 84.5% which outperforming 10 percent than GoogleNet.

Paper Details

Date Published: 6 May 2019
PDF: 7 pages
Proc. SPIE 11069, Tenth International Conference on Graphics and Image Processing (ICGIP 2018), 110692W (6 May 2019); doi: 10.1117/12.2524240
Show Author Affiliations
Jing Ge, Qilu Univ. of Technology (China)
Ying Xie, Qilu Univ. of Technology (China)
Di Wang, Qilu Univ. of Technology (China)


Published in SPIE Proceedings Vol. 11069:
Tenth International Conference on Graphics and Image Processing (ICGIP 2018)
Chunming Li; Hui Yu; Zhigeng Pan; Yifei Pu, Editor(s)

© SPIE. Terms of Use
Back to Top