首页> 外国专利> ARTIFICIAL INTELLIGENCE BASED MACULAR DEGENERATION DISEASE DETERIORATION PREDICTION DEVICES AND METHODS

ARTIFICIAL INTELLIGENCE BASED MACULAR DEGENERATION DISEASE DETERIORATION PREDICTION DEVICES AND METHODS

机译:基于人工智能的黄斑变性疾病劣化预测装置和方法

摘要

To an apparatus for predicting deterioration of macular degeneration based on artificial intelligence, the apparatus for predicting deterioration of macular degeneration based on artificial intelligence includes a database including a plurality of macular degeneration fundus data sets, and a macular degeneration fundus image included in the plurality of macular degeneration fundus data sets. A data preprocessor that performs preprocessing of , constructs an artificial intelligence-based drusen image learning model using a Drusen segmentation mask corresponding to the macular degeneration fundus image, and uses the preprocessed macular degeneration fundus image as the drusen Using the drusen image segmentation unit that generates a drusen region map linked to the macular degeneration fundus image preprocessed by applying to the Zen image learning model, and the macular degeneration fundus image in the macular degeneration fundus image and the drusen image segmentation unit The generated future macular degeneration fundus image is generated by applying the generated drusen region map to the artificial intelligence neural network, and the generated future macular degeneration fundus image is applied to a plurality of artificial intelligence neural networks to determine the authenticity. Includes a prediction image generator can do.
机译:对于基于人工智能预测黄斑变性恶化的装置,该装置基于人工智能预测黄斑变性的恶化的装置包括包括多个黄斑退化基底数据集的数据库,以及包括在多个中的黄斑变性眼底图像黄斑退化基底数据集。执行预处理的数据预处理器构造了使用与黄斑变性眼底图像对应的德鲁森分割掩模构建了基于人工智能的博客图像学习模型,并使用生成的沃定图像分割单元作为司令作为博客使用预处理的黄斑变性眼底图像链接到PURCULAS退化基底图像的德定区域映射通过施加到ZEN图像学习模型,并且在黄斑变性眼底图像中的黄斑变性眼底图像和德鲁森图像分割单元通过施用产生产生的未来的黄斑变性眼底图像将产生的Drusen区域图映射到人工智能神经网络,并且所产生的未来黄斑退化基底图像应用于多个人工智能神经网络以确定真实性。包括预测图像发生器可以做到。

著录项

  • 公开/公告号KR20210153166A

    专利类型

  • 公开/公告日2021-12-17

    原文格式PDF

  • 申请/专利权人 성균관대학교산학협력단;

    申请/专利号KR1020200069350

  • 申请日2020-06-09

  • 分类号G16H50/50;A61B3;A61B3/12;A61B5;G06N3/04;G06N3/08;G06Q10/04;G16H10/60;G16H30/40;

  • 国家 KR

  • 入库时间 2024-06-14 22:33:22

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号