鼻腔内疾患診断支援システムの開発

鵜飼 夏帆; 張 英夏; 向井 信彦; 平野 康次郎; 村上 幸三

doi:10.11371/wiieej.23.04.0_135

Reports of the 308th Technical Conference of the Institute of Image Electronics Engineers of Japan

Session ID : 23-04-40

DOI https://doi.org/10.11371/wiieej.23.04.0_135

Conference information

Host: The Institute of Image Electronics Engineers of Japan

Co-host: The Institute of Image Information and Television Engineers, The Society for Art & Science, The Computer Graphic Arts Society(CG-ARTS)

Name : Reports of the 308th Technical Conference of the Institute of Image Electronics Engineers of Japan

Number : 308

Location : [in Japanese]

Date : March 05, 2024 -

Development of a Diagnostic Support System for Intranasal Disease

*Kaho UKAI, Youngha CHANG, Nobuhiko MUKAI, Kojiro HIRANO, Kouzou MURAKAMI

Author information

Keywords: Intra nasal disease, Endoscope images, Deep learning, Fine tuning, Computer-aided diagnosis

CONFERENCE PROCEEDINGS RESTRICTED ACCESS

Details

Abstract

In this research, a support system has been developed to diagnose whether an endoscopic image shows "severely abnormal nasal cavity (Symptom+)" or "not severely abnormal nasal cavity (Symptom-)". ResNet50 and VGG16 are employed as the deep learning models, and fine-tuning is performed with endoscope images of the nasal cavity after pre-training with ImageNet. The average diagnostic accuracy of stratified 4-fold cross-validation was about 80%, while the recall rate of Symptom+ was about 60%. In the future, we plan to improve the recall rate for practical use.

Corresponding author

Register with J-STAGE for free!