2022 Fiscal Year Final Research Report
Development of a Skin Cancer Diagnosis Support System Based on Highly Accurate Melanoma Structure Pattern Extraction Technology
| Project/Area Number |
19K12054
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 61010:Perceptual information processing-related
|
|---|
| Research Institution | Fukui University of Technology |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2019-04-01 – 2023-03-31
|
| Keywords | ダーモスコピー画像 / メラノーマ構造パターン / mathematical morphology / 深層学習 |
|---|
| Outline of Final Research Achievements |
In this study, a segmentation method for lesion regions, an automatic extraction method for melanoma structural patterns, and an integrated visualization method for multiple structural feature information were developed for dermatoscopic images. The aim was to establish a unified base image processing system for the structural analysis of melanoma, leveraging mathematical morphology-based image processing theory. The research demonstrated that combining the developed morphological image processing techniques with deep learning algorithms improves the accuracy of lesion segmentation, thereby outperforming traditional methods.
|
| Free Research Field |
画像情報解析
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究では，Mathematical morphologyに基づく画像処理手法：Rotational morphological processingによって，ダーモスコピー画像の処理に関する共通基盤を構築した．多様な病変パターンの抽出等において，汎用的かつ論理的一貫性を持つ処理の実現は，実用的意義を有するものである．また，病変領域をセグメンテーションするために考案した深層学習の手法は，ニューラルネットワークが対象のみに着目することを可能にするため，解析対象の認識精度を向上させることができる．幅広い分野での応用展開が期待できる．
|
