WO2020215804A1 - 一种基于深度学习的结肠镜粪便粪水检测方法 - Google Patents

一种基于深度学习的结肠镜粪便粪水检测方法 Download PDF

Info

Publication number
WO2020215804A1
WO2020215804A1 PCT/CN2020/000060 CN2020000060W WO2020215804A1 WO 2020215804 A1 WO2020215804 A1 WO 2020215804A1 CN 2020000060 W CN2020000060 W CN 2020000060W WO 2020215804 A1 WO2020215804 A1 WO 2020215804A1
Authority
WO
WIPO (PCT)
Prior art keywords
feces
fecal
deep learning
image
detection model
Prior art date
Application number
PCT/CN2020/000060
Other languages
English (en)
French (fr)
Inventor
王玉峰
Original Assignee
天津御锦人工智能医疗科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 天津御锦人工智能医疗科技有限公司 filed Critical 天津御锦人工智能医疗科技有限公司
Publication of WO2020215804A1 publication Critical patent/WO2020215804A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/31—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor for the rectum, e.g. proctoscopes, sigmoidoscopes, colonoscopes
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/00002—Operational features of endoscopes
    • A61B1/00004—Operational features of endoscopes characterised by electronic signal processing
    • A61B1/00009—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope
    • A61B1/000096—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope using artificial intelligence
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/00002—Operational features of endoscopes
    • A61B1/00004—Operational features of endoscopes characterised by electronic signal processing
    • A61B1/00009—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/00002—Operational features of endoscopes
    • A61B1/00004—Operational features of endoscopes characterised by electronic signal processing
    • A61B1/00009—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope
    • A61B1/000094—Operational features of endoscopes characterised by electronic signal processing of image signals during a use of endoscope extracting biological structures
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/00002—Operational features of endoscopes
    • A61B1/00043—Operational features of endoscopes provided with output arrangements
    • A61B1/00045—Display arrangement
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/00002—Operational features of endoscopes
    • A61B1/00043—Operational features of endoscopes provided with output arrangements
    • A61B1/00055—Operational features of endoscopes provided with output arrangements for alerting the user
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B1/00—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor
    • A61B1/04—Instruments for performing medical examinations of the interior of cavities or tubes of the body by visual or photographical inspection, e.g. endoscopes; Illuminating arrangements therefor combined with photographic or television appliances
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis

Definitions

  • the invention relates to the technical field of deep learning, in particular to a colonoscope fecal water detection method based on deep learning.
  • the feces and feces inside the intestines are extremely easy to affect the sight of the surgical operator.
  • the doctor must manually flush the feces and feces with water to further inspect the intestines.
  • some doctors may choose not to rinse because they are lazy, and then proceed to the next examination.
  • the colonoscopy quality standard and scoring system if feces or fecal water appears inside the lens, the doctor will be reminded to rinse.
  • the colonoscopy system needs to be able to autonomously identify the feces or feces in the field of view.
  • the existing image processing technology can separate the feces and feces from the image, the recognition rate is not high because the feces of the feces do not have a specific shape and color. And it will cause a great delay to the entire inspection system, resulting in the system not being able to remind the surgeon in real time to flush the feces in the intestinal tract.
  • the purpose of the present invention is to provide a colonoscope fecal water detection method based on deep learning in view of the technical defects existing in the prior art.
  • a colonoscope fecal water detection method based on deep learning including: dividing the video stream from the colonoscope lens in the operating table into two, one part is transmitted to the doctor's operating platform, and the other part of the video stream is preprocessed and sent
  • the fecal and fecal water detection model embedded in the system is used for identification; the fecal and fecal water detection model detects the presence of feces or feces and its probability of occurrence in each frame of image; the information detection results are returned to the doctor's operating platform for display.
  • the stool and fecal water detection model is obtained through the following steps:
  • the feces or fecal water in the image set images are labeled as target detection objects, and the labeled image set is divided into training set and test set:
  • the training set is used to train the formed initialization model, the test set is used to test, and the fecal and fecal water detection model is finally obtained through training and testing.
  • the fecal and fecal water detection model is constructed using the YOLOv3 detection algorithm.
  • the present invention uses a model based on a deep learning algorithm to detect fecal water in a relatively good method, can locate the position and size of the fecal water in the image more quickly and accurately, and solves the problem of using traditional image processing methods to check the The shortcomings that come can generate prompt information for the doctors during the operation in real time.
  • the area proportion of the target object in each image can also be calculated, thereby improving the quality assessment of colonoscopy.
  • Figure 1 is a flowchart of the entire colonoscopy process
  • Figure 2 shows the training process of the entire deep learning model.
  • Fig. 1 is a flow chart of the entire colonoscopy procedure. As shown in Fig. 1, the steps of a colonoscope fecal water detection method based on deep learning of the present invention are as follows:
  • Step 1 When the operation starts, the video stream from the colonoscopy lens in the operating table is divided into two, one part is transmitted to the doctor's operating platform, and the other part is sent to the deep learning model built on the deep learning network-feces Water detection model (embedded in the artificial intelligence detection module of the system) for identification;
  • Step 2 Preprocess the video stream, and then send it to the deep learning network for recognition, and call the fecal and fecal water detection model for each frame of the video stream. Detect whether there is feces or feces in each frame of image and the probability that the detected target is feces or feces;
  • Step 3 Return the test result of step 2 to the doctor's operating platform for display, prompting the doctor for further processing;
  • the overall flow chart is shown in Figure 1.
  • the stool and fecal water detection model based on the deep learning network in step 2 is developed using the YOLOv3 algorithm to meet the real-time requirements of the entire colonoscopy.
  • Figure 2 shows the training process of the entire deep learning model (ie, fecal water detection model) as follows:
  • Step 1 Obtain a clear and bright image collection with fecal and fecal water intercepted during colonoscopy from the hospital database;
  • Step 2 Perform target labeling on the image set obtained in step 1, and label the target objects such as feces and water in the image using labelimg.
  • the labeled image set is divided into two parts: training set and test set, which is convenient for training and testing, including:
  • Step 2.1 uniformly crop the image set obtained in step 1, and crop it into an image set of uniform size and same format
  • Step 2.2 Use labelimg to label the target in the image collection to obtain a complete fecal and water image collection;
  • Step 2.3 Select 1500 images in the complete image set as the training set and 300 images as the test set;
  • Step 3 Input the selected training set into the initialization model of YOLOv3, set the parameters in the training process, and then perform training.
  • Step 4. Save the learning network obtained after the training.
  • Step 5 Transmit the test set as input to the final deep learning network, and detect the output result of the deep learning network.
  • Step 6 Embed the trained stool and fecal water detection model based on the deep learning network into the colonoscopy quality assessment system.
  • Step 7 check the results output by the fecal and fecal water detection model by the evaluation system to determine whether there is feces or feces in the current image.
  • Step 8 By calculating the output position information of the target detection object, the area ratio of the target detection object in the current image can be obtained.
  • the present invention uses a model based on a deep learning algorithm to detect fecal water in a relatively good method, can locate the position and size of the fecal water in the image more quickly and accurately, and solves the problem of using traditional image processing methods to check the The shortcomings that come can generate prompt information for the doctors during the operation in real time.
  • the area proportion of the target object in each image can also be calculated, thereby improving the quality assessment of colonoscopy.

Abstract

一种基于深度学习的结肠镜粪便粪水检测方法，包括：将手术台中肠镜镜头传出的视频流一分为二，一部分传输到医生的操作平台上，另一部分视频流进行预处理后送到嵌入到系统中的粪便粪水检测模型进行识别；粪便粪水检测模型对每一帧图像出现粪便或粪水及其出现概率进行检测；将检测结果返回到医生操作平台显示。该方法可以更快速、更准确的定位出图像中粪便粪水的位置及大小，解决了利用传统图像处理的方法检查所带来的问题，可以实时的为手术过程中的医生产生提示信息。

Description

一种基于深度学习的结肠镜粪便粪水检测方法 技术领域
本发明涉及深度学习技术领域，特别是涉及一种基于深度学习的结肠镜粪便粪水检测方法。
背景技术
在结肠镜检查过程中，位于肠道内部的粪便粪水极其容易影响到手术操作人员的视线，这时医生就要手动的用水将粪水粪便冲洗掉，从而进一步的对肠道内进行检查。部分医生在手术过程中可能会因为偷懒而选择不进行冲洗，进而进行接下来的检查。为了防止这样的情况发生，在肠镜检查质量标准及打分系统中，若镜头内部出现了粪便或粪水，将会提醒医生进行冲洗操作。
若要想对医生提示冲洗操作，这就需要肠镜检查系统可以自主的识别出视野内的粪水或是粪便。现有的图像处理的技术虽然可以将粪便粪水从图像中分离出来，但是由于粪水粪便自身没有一个特定的形状及颜色等原因，识别率并不高。并且会对整个检查系统造成很大的延迟，导致系统无法实时的提醒手术医生及时的去冲洗肠道内的粪便粪水。
发明内容
本发明的目的是针对现有技术中存在的技术缺陷，而提供一种基于深度学习的结肠镜粪便粪水检测方法。
为实现本发明的目的所采用的技术方案是：
一种基于深度学习的结肠镜粪便粪水检测方法，包括：将手术台中肠镜镜头传出的视频流一分为二，一部分传输到医生的操作平台上，另一部分视频流进行预处理后送到嵌入到系统中的粪便粪水检测模型进行识别；粪便粪水检测模型对每一帧图像出现粪便或粪水及其出现概率进行检测；将息检测结果返回 到医生操作平台显示。
其中，所述粪便粪水检测模型通过以下步骤而获得：
从医院数据库中获取在肠镜检查过程中截取的清晰的带有粪便或粪水的图像集；
将图像集图像中粪便或粪水作为目标检测物标注，将标注好后的图像集分为训练集和测试集：
利用训练集对形成的初始化模型中进行训练，利用测试集进行测试，最终经训练测试而获得所述粪便粪水检测模型。
其中，所述粪便粪水检测模型利用YOLOv3检测算法构建。
与现有技术相比，本发明的有益效果是：
本发明利用基于深度学习算法的模型来检测粪便粪水的方法相对较好，可以更快速更准确的定位出图像中粪便粪水的位置及大小，解决的了利用传统图像处理的方法检查所带来的缺点，可实时的为手术过程中的医生产生提示信息。
另外，利用学习网络输出的粪便粪水的位置信息，还可以计算出每张图像中目标物体的面积占比，从而完善肠镜检查质量评定。
附图说明
图1是整个肠镜检查过程的流程图；
图2为整个深度学习模型的训练过程。
具体实施方式
以下结合附图和具体实施例对本发明作进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
图1是整个肠镜检查过程中的流程图，如图1所示，本发明的一种基于深度学习的结肠镜粪便粪水检测方法步骤如下：
步骤1：当手术开始时，将手术台中肠镜镜头传出的视频流一分为二，一部分传输到医生的操作平台上，另一部分送到基于深度学习网络所构建的深度学习模型-粪便粪水检测模型(嵌入在系统的人工智能检测模块中)进行识别；
步骤2：对视频流进行预处理，然后送到深度学习网络中进行识别，对视频流中的每一帧图像调用粪便粪水检测模型。检测每一帧图像中是否出现粪便或是粪水及检测到的目标为粪便或粪水的概率；
步骤3：将步骤2的检测结果返回到医生操作平台上进行显示，提示医生进行进一步处理；
整体的流程图如图1，步骤2中的基于深度学习网络构建的粪便粪水检测模型，使用YOLOv3算法开发，以满足整个肠镜检查中所需要的实时性。
图2为整个深度学习模型(即粪便粪水检测模型)的训练过程如下：
步骤1、从医院数据库中获取在肠镜检查过程中截取的清晰明亮的带有粪便粪水的图像集；
步骤2、将步骤1中得到的图像集进行目标标注，使用labelimg将图像中的粪便粪水等目标物进行标注。将标注好后的图像集分为训练集和测试集两个部分，便于训练及测试，具体包括：
步骤2.1、将步骤1得到的图像集进行统一裁剪，裁剪为统一大小，相同格式的图像集；
步骤2.2、使用labelimg将图像集中的目标物进行标注，得到完整的粪便粪水图像集；
步骤2.3、挑选完整图像集中的1500张作为训练集，300张作为测试集；
步骤3、将挑选出的训练集输入到YOLOv3的初始化模型中，设定好训练过程中的参数，然后进行训练。
步骤4、将训练结束后得到的学习网络进行保存。
步骤5、将测试集作为输入传输到最终的深度学习网络中，检测深度学习网络输出的结果。
步骤6、将训练好的基于深度学习网络的粪便粪水检测模型嵌入到肠镜检查质量评定系统中。
步骤7、手术过程中，通过评定系统对粪便粪水检测模型输出的结果检查，即可判断当前图像中是否有粪便或者粪水。
步骤8、通过对输出的目标检测物的位置信息进行计算，从而可以得到目标检测物在当前图像中所占面积比例。
与现有技术相比，本发明的有益效果是：
本发明利用基于深度学习算法的模型来检测粪便粪水的方法相对较好，可以更快速更准确的定位出图像中粪便粪水的位置及大小，解决的了利用传统图像处理的方法检查所带来的缺点，可实时的为手术过程中的医生产生提示信息。
另外，利用学习网络输出的粪便粪水的位置信息，还可以计算出每张图像中目标物体的面积占比，从而完善肠镜检查质量评定。
以上所述仅是本发明的优选实施方式，应当指出的是，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (3)

  1. 一种基于深度学习的结肠镜粪便粪水检测方法，其特征在于，包括：
    将手术台中肠镜镜头传出的视频流一分为二，一部分传输到医生的操作平台上，另一部分视频流进行预处理后送到嵌入到系统中的粪便粪水检测模型进行识别；粪便粪水检测模型对每一帧图像出现粪便或粪水及其出现概率进行检测；将息检测结果返回到医生操作平台显示。
  2. 如权利要求1所述基于深度学习的结肠镜粪便粪水检测方法，其特征在于，所述粪便粪水检测模型通过以下步骤而获得：
    从医院数据库中获取在肠镜检查过程中截取的清晰的带有粪便或粪水的图像集；
    将图像集图像中粪便或粪水作为目标检测物标注，将标注好后的图像集分为训练集和测试集：
    利用训练集对形成的初始化模型中进行训练，利用测试集进行测试，最终经训练测试而获得所述粪便粪水检测模型。
  3. 如权利要求1所述基于深度学习的结肠镜粪便粪水检测方法，其特征在于，所述粪便粪水检测模型利用YOLOv3检测算法构建。
PCT/CN2020/000060 2019-04-25 2020-04-09 一种基于深度学习的结肠镜粪便粪水检测方法 WO2020215804A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910340439.1A CN111839446A (zh) 2019-04-25 2019-04-25 一种基于深度学习的结肠镜粪便粪水检测方法
CN201910340439.1 2019-04-25

Publications (1)

Publication Number Publication Date
WO2020215804A1 true WO2020215804A1 (zh) 2020-10-29

Family

ID=72941012

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/000060 WO2020215804A1 (zh) 2019-04-25 2020-04-09 一种基于深度学习的结肠镜粪便粪水检测方法

Country Status (2)

Country Link
CN (1) CN111839446A (zh)
WO (1) WO2020215804A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183674A (zh) * 2020-11-06 2021-01-05 南昌航空大学 一种粪便宏观图像颜色和性状多任务识别方法及系统

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010057541A (ja) * 2008-09-01 2010-03-18 Tooru Takita 医療用検査装置
CN106651883A (zh) * 2016-12-30 2017-05-10 四川沃文特生物技术有限公司 基于机器视觉的粪便形态识别方法
CN106682633A (zh) * 2016-12-30 2017-05-17 四川沃文特生物技术有限公司 基于机器视觉的粪便镜检图像有形成分的分类识别方法
CN108695001A (zh) * 2018-07-16 2018-10-23 武汉大学人民医院（湖北省人民医院） 一种基于深度学习的癌症病灶范围预测辅助系统及方法
CN109063747A (zh) * 2018-07-16 2018-12-21 武汉大学人民医院（湖北省人民医院） 基于深度学习的肠道病理切片图像识别分析系统及方法
CN109447985A (zh) * 2018-11-16 2019-03-08 青岛美迪康数字工程有限公司 结肠镜图像分析方法、装置及可读存储介质
CN109598704A (zh) * 2018-11-19 2019-04-09 电子科技大学 一种基于bp神经网络的粪便显微图像清晰度评价方法
CN109615633A (zh) * 2018-11-28 2019-04-12 武汉大学人民医院（湖北省人民医院） 一种基于深度学习的肠镜下克罗恩病辅助诊断系统及方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011156001A1 (en) * 2010-06-07 2011-12-15 Sti Medical Systems, Llc Versatile video interpretation,visualization, and management system
CN109255044A (zh) * 2018-08-31 2019-01-22 江苏大学 一种基于YOLOv3深度学习网络的图像智能标注方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010057541A (ja) * 2008-09-01 2010-03-18 Tooru Takita 医療用検査装置
CN106651883A (zh) * 2016-12-30 2017-05-10 四川沃文特生物技术有限公司 基于机器视觉的粪便形态识别方法
CN106682633A (zh) * 2016-12-30 2017-05-17 四川沃文特生物技术有限公司 基于机器视觉的粪便镜检图像有形成分的分类识别方法
CN108695001A (zh) * 2018-07-16 2018-10-23 武汉大学人民医院（湖北省人民医院） 一种基于深度学习的癌症病灶范围预测辅助系统及方法
CN109063747A (zh) * 2018-07-16 2018-12-21 武汉大学人民医院（湖北省人民医院） 基于深度学习的肠道病理切片图像识别分析系统及方法
CN109447985A (zh) * 2018-11-16 2019-03-08 青岛美迪康数字工程有限公司 结肠镜图像分析方法、装置及可读存储介质
CN109598704A (zh) * 2018-11-19 2019-04-09 电子科技大学 一种基于bp神经网络的粪便显微图像清晰度评价方法
CN109615633A (zh) * 2018-11-28 2019-04-12 武汉大学人民医院（湖北省人民医院） 一种基于深度学习的肠镜下克罗恩病辅助诊断系统及方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
吴康英 等 (WU, KANGYING ET AL.): "纤维结肠镜检查术前肠道清洁的体会 (non-official translation: Experience of Preoperative Colon Cleansing prior to Colon Fiberscopy)", 广东医学院学报 (JOURNAL OF GUANGDONG MEDICAL COLLEGE), vol. 19, no. 4, 31 August 2001 (2001-08-31), XP55746530, ISSN: 1005-4057, DOI: 20200603165841Y *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183674A (zh) * 2020-11-06 2021-01-05 南昌航空大学 一种粪便宏观图像颜色和性状多任务识别方法及系统
CN112183674B (zh) * 2020-11-06 2022-06-10 南昌航空大学 一种粪便宏观图像颜色和性状多任务识别方法及系统

Also Published As

Publication number Publication date
CN111839446A (zh) 2020-10-30

Similar Documents

Publication Publication Date Title
CN103442628B (zh) 用于检查宫颈的医疗器械
TWI684997B (zh) 骨齡評估與身高預測模型之建立方法、骨齡評估與身高預測系統及骨齡評估與身高預測方法
WO2020215810A1 (zh) 一种基于图像识别的结肠镜手术中窄带成像检测方法
CN109102491A (zh) 一种胃镜图像自动采集系统及方法
Wannous et al. Supervised tissue classification from color images for a complete wound assessment tool
WO2020215807A1 (zh) 一种基于深度学习提高结肠镜腺瘤性息肉检出率的方法
Hatanaka et al. Vertical cup-to-disc ratio measurement for diagnosis of glaucoma on fundus images
WO2007053470A3 (en) Detecting humans via their pose
EP3961644B1 (en) Image recognition based workstation for evaluation on quality check of colonoscopy
US20230270404A1 (en) Carotid artery ultrasonic examination report generation system based on multi-modal information
CN110363768A (zh) 一种基于深度学习的早期癌病灶范围预测辅助系统
CN111080639A (zh) 基于人工智能的多场景消化道内镜图像识别方法及系统
CN113139962B (zh) 用于脊柱侧凸概率评估的系统和方法
CN111462049A (zh) 一种乳腺超声造影视频中病灶区形态自动标注方法
WO2020215804A1 (zh) 一种基于深度学习的结肠镜粪便粪水检测方法
WO2023143014A1 (zh) 一种基于人工智能的内窥镜辅助检查方法及装置
CN108510461B (zh) 耳膜状态实时分析方法
CN116109812A (zh) 一种基于非极大值抑制阈值优化的目标检测方法
CN113017702A (zh) 超声内镜小探头伸出长度的识别方法、系统及存储介质
CN115345858A (zh) 一种超声心动图质量评价方法、系统及终端设备
CN111402231A (zh) 一种用于肺部ct影像质量的自动评估系统及方法
CN108074228B (zh) 耳膜状态实时分析系统
CN115661037A (zh) 一种胶囊内镜辅助检测方法、装置、系统、设备及介质
CN109028234B (zh) 一种能够对烟雾等级进行标识的油烟机
Pan et al. Bleeding detection from wireless capsule endoscopy images using improved euler distance in CIELab

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20794312

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20794312

Country of ref document: EP

Kind code of ref document: A1