Label Generation Network based on Self-selected Historical Information for Multiple Disease Classification on Chest Radiography

Yuelin Hu; Yuejie Zhang; Tao Zhang; Shang Gao; Weiguo Fan

doi:10.1109/BIBM49941.2020.9313507

Back

Conference proceeding

Label Generation Network based on Self-selected Historical Information for Multiple Disease Classification on Chest Radiography

Yuelin Hu, Yuejie Zhang, Tao Zhang, Shang Gao and Weiguo Fan

2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp.1015-1019

12/16/2020

DOI: 10.1109/BIBM49941.2020.9313507

View Online

Abstract

Deep learning has made significant break through's in image classification, but accurate diagnosis on chest radiography remains challenging due to a variety of potential diseases contained in one scan. Complex relations among diseases have significant clinical meanings, but are always ignored in most of previous work. Thus in this paper, we propose a novel Label Generation Network (LGN) which treats the label sequence as the caption of a radiology image and utilizes RNN to generate the disease labels according to the semantic relations and co-occurrence dependency among them. However, the sequential generation process of RNN makes it hard to capture the complex topological relations among diseases. To mitigate this problem, a Historical Information Module (HIM) is especially introduced to LGN, in which all the generated labels are fully considered when generating a new label. Moreover, a specific self-attention mechanism is applied in HIM to learn the topological disease relations and utilize them to select useful historical information which can provide positive guidance to the prediction of new label. Very positive results have been obtained in our experiments on the benchmark dataset of Chest X-ray14, which significantly outperform the state-of-the-art methods.

Chest radiography

Computational modeling

Diagnostic radiography

Diseases

Feature extraction

Generators

Historical information

Label Generation Network

Multi-label classification

Self-attention mechanism

Training

X-ray imaging

Details

Title: Subtitle: Label Generation Network based on Self-selected Historical Information for Multiple Disease Classification on Chest Radiography
Creators: Yuelin Hu - Fudan University
Yuejie Zhang - Fudan University
Tao Zhang - Shanghai University of Finance and Economics
Shang Gao - Deakin University
Weiguo Fan - University of Iowa
Resource Type: Conference proceeding
Publication Details: 2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp.1015-1019
Publisher: IEEE
DOI: 10.1109/BIBM49941.2020.9313507
ISSN: 2156-1125
eISSN: 2156-1133
Grant note: Ministry of Education (10.13039/501100002701)
Language: English
Date published: 12/16/2020
Academic Unit: Business Analytics
Record Identifier: 9984380419802771

Metrics

3 Record Views