Development of advanced CAD system for mammography and its practical application

先进乳腺X线摄影CAD系统的开发及其实际应用

• 批准号: 10555136
• 负责人: KOBATAKE Hidefumi
• 金额: $ 7.81万
• 依托单位: Tokyo University of Agriculture and Technology
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B).
• 财政年份: 1998
• 资助国家: 日本
• 起止时间: 1998 至 2000
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-10555136/
• 关键词: Computer Aided Diagnosis; CAD; Mammography; breast cancer; microcalcifications; tumor; 腫瘤影検出; 診断支援システム

The purpose of this research project is to develop an improved CAD system for mammography and to heighten the system performance. The main results are as follows.[1]System robust to variations of imaging conditionThe conventional CAD system has been installed is five medical institutions in Japan and its performance had been tested . Each medical institution has its own imaging conditions and the performance of the conventional system depends on them. We have developed new processing algorithm, which makes the CAD system insensitive to the variation of imaging conditions.[2]Revised tumor detection methodAdaptive ring filter has been developed, which makes it possible to detect tumors with rough and irregular density distributions. The performance of the adaptive ring filter has been shown to be superior to that of the iris filter.[3]Detection of tumors close to or attached to thoracic wallTumors close to or attached to thoracic wall tend to be missed by the conventional CAD system. The detection algorithm for them has been revised. That is, the candidates in the area near thoracic wall are recognized by the different algorithm developed by this project. The performance of the new system has become higher than that of the conventional system.[4]Advanced microcalcification detection methodCoarse-calcification causes false microcalcifications. We have developed a new method for coarse-calcification detection, which can help decreasing false microcalcification clusters. Complicated background can cause false microcalcifications. We have also developed an algorithm to exclude them.These methods described above have been installed in the new CAD system.

本研究的目的是开发一种改进的乳腺x线造影CAD系统，以提高系统的性能。主要结果如下：传统的CAD系统已在日本的五家医疗机构安装，并进行了性能测试。每个医疗机构都有自己的成像条件，常规系统的性能取决于这些条件。我们开发了新的处理算法，使CAD系统对成像条件的变化不敏感。改进了肿瘤检测方法自适应环滤波器的发展，使密度分布粗糙、不规则的肿瘤检测成为可能。自适应环形滤波器的性能优于虹膜滤波器。[3]靠近或附着胸壁肿瘤的检测靠近或附着胸壁的肿瘤在传统的CAD系统中容易被遗漏。对它们的检测算法进行了修改。即在胸壁附近区域的候选者被本项目开发的不同算法识别。新系统的性能比传统系统有所提高。高级微钙化检测方法粗钙化引起假微钙化。我们开发了一种新的粗钙化检测方法，可以帮助减少假微钙化簇。复杂的背景可引起假的微钙化。我们还开发了一种算法来排除它们。上述方法已应用于新的CAD系统中。

项目成果

H.Kobatake, H.Takeo and S.Nawano: "Microcalcification Detection System for Full-Digital Mammography"Digital Mammography. 201-204 (1998)

H.Kobatake、H.Takeo 和 S.Nawano：“全数字乳房 X 线摄影的微钙化检测系统”数字乳房 X 线摄影。

M.Murakami, H.Kobatake, H.Takeo and S.Nawano: "Feature Extraction of Malignant Tumors on Mammograms"Medical Imaging Technology. Vol.17, No.6. 681-689 (1999)

M.Murakami、H.Kobatake、H.Takeo 和 S.Nawano：“乳房 X 光照片上恶性肿瘤的特征提取”医学成像技术。

村上政行: "乳房X線上の悪性腫瘍瘤影の特徴抽出"Medical Imaging Technology. 17・6. 681-689 (1999)

村上雅之：“乳腺X射线恶性肿瘤阴影的特征提取”医学成像技术17・6（1999）。

橋本磁: "左右マンモグラム定性的特徴量の比較に基づく偽陽性陰影の削減"電子情報通信学会論文誌. J-83-D-II・1. 415-418 (2000)

Masaru Hashimoto：“基于左右乳房 X 光检查的定性特征的减少”，电子、信息和通信工程师学会会刊 J-83-D-II·1（2000 年）。

J.Wei, Y.Hagihara and H.Kobatake: "Convergence Index Filter for Detection of Lung Nodule Candidates"The Transactions of Institute of Electrics, Information and Communication Engineers, Vol.J83-D・II, No.1. 118-125 (2000)

J.Wei、Y.Hagihara 和 H.Kobatake：“用于检测肺结节候选者的收敛指数过滤器”电气、信息和通信工程师学会学报，第 J83-D・II 卷，第 118-125 期。 (2000)