WO2019223148A1 - 基于面部识别的痤疮判断方法、终端及存储介质 - Google Patents

基于面部识别的痤疮判断方法、终端及存储介质 Download PDF

Info

Publication number
WO2019223148A1
WO2019223148A1 PCT/CN2018/102201 CN2018102201W WO2019223148A1 WO 2019223148 A1 WO2019223148 A1 WO 2019223148A1 CN 2018102201 W CN2018102201 W CN 2018102201W WO 2019223148 A1 WO2019223148 A1 WO 2019223148A1
Authority
WO
WIPO (PCT)
Prior art keywords
deep learning
image
learning model
facial
trained deep
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2018/102201
Other languages
English (en)
French (fr)
Inventor
王健宗
吴天博
黄章成
肖京
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ping An Technology Shenzhen Co Ltd
Original Assignee
Ping An Technology Shenzhen Co Ltd
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 Ping An Technology Shenzhen Co Ltd filed Critical Ping An Technology Shenzhen Co Ltd
Publication of WO2019223148A1 publication Critical patent/WO2019223148A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30088—Skin; Dermal

Definitions

  • the present application relates to the field of facial recognition, and in particular, to an acne judgment method, terminal, and storage medium based on facial recognition.
  • Acne is a common chronic inflammation of the hair follicles and sebaceous glands caused by many factors.
  • about 85% of the population develops acne between the ages of 12-24, of which only 10% -30% of acne patients are treated in dermatology, but there are only about 22,000 dermatologists in China, and the number of doctors has increased.
  • the uneven distribution of resources makes it difficult for many acne patients to obtain timely and effective treatment in the early stages of the disease, eventually leading to worsening of symptoms and even serious psychological problems such as anxiety and depression.
  • this application proposes a facial recognition-based acne determination method, terminal, and storage medium.
  • the present application proposes a terminal.
  • the terminal includes a memory and a processor.
  • the memory stores a facial recognition-based acne judgment program that can be run on the processor.
  • the identified acne judging program is executed by the processor, the following steps are implemented: obtaining a trained deep learning model; loading the trained deep learning model; collecting a facial image; and according to the collected facial image and the trained image Deep learning model for image recognition; obtaining image recognition results, and uploading the image recognition results to a cloud pathology database, the cloud pathology database being used for pathological retrieval based on the image recognition results; and receiving the cloud pathology database
  • the results of pathological retrieval are presented to the user.
  • the present application also provides a facial recognition-based acne determination method.
  • the facial recognition-based acne determination method includes: acquiring a trained deep learning model; loading the trained deep learning model; Collect facial images; perform image recognition according to the acquired facial images and the trained deep learning model; acquire image recognition results, and upload the image recognition results to a cloud pathology database, which is used to Performing pathological retrieval on the image recognition result; and receiving the pathological retrieval result of the cloud pathological database and presenting it to the user.
  • the present application further provides a storage medium storing a facial recognition-based acne determination program, and the facial recognition-based acne determination program may be executed by at least one processor so that The at least one processor executes the steps of the facial recognition-based acne determination method as described above.
  • FIG. 1 is a schematic diagram of a hardware structure of a terminal that implements various embodiments of the present application
  • FIG. 2 is an application environment diagram of a terminal of the present application
  • FIG. 3 is a program module diagram of a first embodiment of a facial recognition-based acne determination program of the present application
  • FIG. 4 is a schematic diagram of an implementation process of a first embodiment of an acne determination method based on facial recognition of the present application
  • the terminal can be implemented in various forms.
  • the terminals described in this application may include mobile phones, tablets, laptops, palmtop computers, Personal Digital Assistants (PDAs), Portable Media Players (PMPs), navigation devices, Wearable devices, smart bracelets, pedometers and other terminals that can have image acquisition functions, and fixed terminals such as digital TVs, desktop computers and other devices that can have image acquisition functions.
  • PDAs Personal Digital Assistants
  • PMPs Portable Media Players
  • navigation devices wearable devices
  • smart bracelets smart bracelets
  • pedometers and other terminals that can have image acquisition functions
  • fixed terminals such as digital TVs, desktop computers and other devices that can have image acquisition functions.
  • a mobile terminal will be taken as an example for explanation.
  • the configuration according to the embodiment of the present application can also be applied to a fixed type terminal.
  • FIG. 1 is a schematic diagram of a hardware structure of a terminal that implements various embodiments of the present application.
  • the terminal 100 may include an RF (Radio Frequency) unit 101, a WiFi module 102, an audio output unit 103, and an A / V. (Audio / Video)
  • the terminal device may include more or less components than shown in the figure, or some components may be combined, or different component arrangements may be arranged. .
  • the memory 109 may be used to store software programs and various data.
  • the memory 109 may mainly include a storage program area and a storage data area, where the storage program area may store an operating system, at least one application required by a function (such as a sound playback function, an image playback function, etc.), etc .; the storage data area may store data according to Data (such as audio data, phone book, etc.) created by the use of mobile phones.
  • the memory 109 may include a high-speed random access memory, and may further include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.
  • the terminal 100 may further include a Bluetooth module and the like, and details are not described herein again.
  • FIG. 2 is an application environment diagram of a terminal of the present application. It is based on the terminal 100 shown in FIG. 1, by collecting the target image of the target object 400 and performing a certain preliminary processing on the collected image, the trained deep learning model is imported for calculation, the calculation result is obtained, and the calculation result is uploaded To the cloud pathology database 500, the cloud pathology database 500 is configured to perform pathological retrieval according to the image recognition result; and then receive the pathological retrieval result of the cloud pathology database 500 and present it to the user.
  • deep learning can be used to accurately and quickly identify symptoms, combined with a pathological database for fast pathological retrieval, and then for patients or doctors to make reference judgments, which improves the efficiency of acne treatment.
  • the present application proposes an acne determination program 300 based on facial recognition.
  • FIG. 3 it is a program module diagram of the first embodiment of the acne determination program 300 based on facial recognition in the present application.
  • the facial recognition-based acne determination program 300 includes a series of computer program instructions stored on the memory 109 in the terminal 100. When the computer program instructions are executed by the processor 110, the present application can be implemented. The facial recognition-based acne judgment operation of each embodiment.
  • the facial recognition-based acne determination program 300 may be divided into one or more modules based on specific operations implemented by the computer program instructions. For example, in FIG. 3, the facial recognition-based acne determination program 300 may be divided into a training module 301, a loading module 302, a collection module 303, a recognition module 304, a retrieval module 305, and an output module 306. among them:
  • the training module 301 is configured to obtain a trained deep learning model.
  • the deep learning model may be a VGG-16 model or an inceptionV3 model.
  • Both of the above deep learning models are image classification models, and are deep learning model codes that are open on Google ’s official website. After downloading the VGG-16 model or inceptionV3 model code, the training data is used to train the above VGG-16 model or inceptionV3 model, and finally a trained deep learning model is obtained.
  • the step of acquiring the trained deep learning model specifically includes:
  • acne can be divided into 8 categories according to clinical manifestations: acne acne, pimples acne, pustular acne, cystic acne, nodular acne, atrophic acne, polymerized acne, and cachexia acne.
  • the transfer learning method is used to load the training data into a deep learning model for training to obtain a trained deep learning model, that is, the deep learning model is used to iteratively calculate a predetermined number of times for the loaded training data to obtain the trained training data. Deep learning models.
  • the loading module 302 is configured to load the trained deep learning model.
  • the terminal includes a tensorflow module, wherein the tensorflow is a second-generation artificial intelligence learning system developed by Google based on DistBelief, which can support deep neural learning networks for fast calculation of flow graphs.
  • the deep learning model is also initialized to prepare for subsequent calculation of the deep learning model.
  • the acquisition module 303 is configured to acquire a facial image.
  • the acquisition of a facial image is mainly performed by calling an API interface of an image acquisition device of a terminal, and then starting the image acquisition device through the API interface and receiving a facial image collected by the image acquisition device.
  • the image acquisition device is preferably a camera.
  • the acquisition module 303 first prompts the user to choose to upload photos or take photos through the terminal display interface; if the user chooses to upload photos, the user is prompted to upload positive, left, and right facial photos. One is used as the facial image; if the user chooses to take a photo, the user is prompted to expose the face to a natural light source, and one each of the positive, left, and right facial photos is taken as the facial image.
  • the collecting module 303 may further preprocess the facial image, that is, crop the captured facial image with a preset size, and compress the cropped facial image. Because the facial images generally collected are more than 1M in size, too large pictures have an impact on the data transmission efficiency in the calculation process. Therefore, the efficiency of the above deep learning model calculation is improved by cutting and compressing in advance.
  • the recognition module 304 is configured to perform image recognition according to the collected facial image and the trained deep learning model.
  • the recognition module 304 sends the collected facial image to the aforementioned tensorflow module, and the tensorflow module calculates the facial image in the trained deep learning model.
  • the essence of any image is a digital array of a certain size for a computer.
  • the retrieval module 305 is configured to obtain an image recognition result, and upload the image recognition result to a cloud pathology database 500.
  • the cloud pathology database 500 is configured to perform pathological retrieval according to the image recognition result.
  • the image recognition result is a class represented by the output port with the largest value among the 32 output results, and the result is uploaded to the cloud pathology data 500.
  • the cloud pathology database 500 stores the corresponding diseases. Symptoms, treatment methods, precautions and other pathologies. For example, pictures of symptoms corresponding to Class I acne, how to treat, precautions during treatment and rehabilitation, and so on.
  • the output module 306 is configured to receive the pathological retrieval result of the cloud pathological database and present it to the user.
  • the output module 306 can output the matching result through the display unit 106 of the terminal 100, thereby providing the user with an intuitive preliminary diagnosis of the condition.
  • the facial recognition-based acne judgment program 300 proposed in the present application can quickly obtain the corresponding diagnosis results and treatment information after identifying the collected facial images, thereby providing patients with more efficient treatment plan.
  • this application also proposes a method for determining acne based on facial recognition.
  • FIG. 4 is a schematic diagram of an implementation process of a first embodiment of an acne determination method based on facial recognition in the present application.
  • Step S401 Obtain a trained deep learning model.
  • the deep learning model may be a VGG-16 model or an inceptionV3 model.
  • Both of the above deep learning models are image classification models, and are deep learning model codes that are open on Google ’s official website. After downloading the VGG-16 model or inceptionV3 model code, the training data is used to train the above VGG-16 model or inceptionV3 model, and finally a trained deep learning model is obtained.
  • the step of acquiring the trained deep learning model specifically includes:
  • acne can be divided into 8 categories according to clinical manifestations: acne acne, pimples acne, pustular acne, cystic acne, nodular acne, atrophic acne, polymerized acne, and cachexia acne.
  • the transfer learning method is used to load the training data into a deep learning model for training to obtain a trained deep learning model, that is, the deep learning model is used to iteratively calculate a predetermined number of times for the loaded training data to obtain the trained Deep learning models.
  • Step S402 Load the trained deep learning model.
  • the terminal includes a tensorflow module, wherein the tensorflow is a second-generation artificial intelligence learning system developed by Google based on DistBelief, which can support deep neural learning networks for fast calculation of flow graphs.
  • the deep learning model is also initialized to prepare for subsequent calculation of the deep learning model.
  • step S403 a facial image is collected.
  • the acquisition of a facial image is mainly performed by calling an API interface of an image acquisition device of a terminal, and then starting the image acquisition device through the API interface and receiving a facial image collected by the image acquisition device.
  • the image acquisition device is preferably a camera.
  • the acquisition module 303 first prompts the user to choose to upload photos or take photos through the terminal display interface; if the user chooses to upload photos, prompts the user to upload positive, left, and right facial photos One is used as the facial image; if the user chooses to take a photo, the user is prompted to expose the face to a natural light source, and one each of the positive, left, and right facial photos is taken as the facial image.
  • the collecting module 303 may further preprocess the facial image, that is, crop the captured facial image with a preset size, and compress the cropped facial image. Because the facial images generally collected are more than 1M in size, too large pictures have an impact on the data transmission efficiency in the calculation process. Therefore, the efficiency of the above deep learning model calculation is improved by cutting and compressing in advance.
  • step S404 image recognition is performed according to the collected facial image and the trained deep learning model.
  • the recognition module 304 sends the collected facial image to the aforementioned tensorflow module, and the tensorflow module calculates the facial image in the trained deep learning model.
  • the essence of any image is a digital array of a certain size for a computer.
  • Step S405 Obtain an image recognition result, and upload the image recognition result to a cloud pathology database 500, where the cloud pathology database 500 is used for pathological retrieval according to the image recognition result.
  • the image recognition result is a class represented by the output port with the largest value among the 32 output results, and the result is uploaded to the cloud pathology data 500.
  • the cloud pathology database 500 stores the corresponding diseases. Symptoms, treatment methods, precautions and other pathologies. For example, pictures of symptoms corresponding to Class I acne, how to treat, precautions during treatment and rehabilitation, and so on.
  • step S406 a pathological retrieval result of the cloud pathological database is received and presented to the user.
  • the display unit 106 of the terminal 100 is used to output the matching result, thereby providing the user with an intuitive preliminary diagnosis of the condition.
  • the facial recognition-based acne judgment method proposed in this application can quickly obtain the corresponding diagnosis results and treatment information after identifying the collected facial images, thereby providing patients with more efficient treatment Program.
  • This application also provides another implementation manner, that is, providing a storage medium storing a facial recognition-based acne determination program, and the facial recognition-based acne determination program may be executed by at least one processor to The at least one processor is caused to execute the steps of the facial recognition-based acne determination method as described above.
  • the methods in the above embodiments can be implemented by means of software plus a necessary universal hardware platform, and of course, also by hardware, but in many cases the former is better.
  • Implementation Based on such an understanding, the technical solution of this application that is essentially or contributes to the existing technology can be embodied in the form of a software product that is stored in a storage medium (such as ROM / RAM, magnetic disk, The optical disc) includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in the embodiments of the present application.
  • a terminal device which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Medical Informatics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Biomedical Technology (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Quality & Reliability (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Radiology & Medical Imaging (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Evolutionary Biology (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

一种基于面部识别的痤疮判断方法，应用于终端，包括：获取训练后的深度学习模型（S401）；加载所述训练后的深度学习模型（S402）；采集面部图像（S403）；根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别（S404）；获取图像识别结果，并将所述图像识别结果上传至云端病理数据库，所述云端病理数据库用于根据所述图像识别结果进行病理检索（S405）；接收所述云端病理数据库的病理检索结果并呈现给用户（S406）。所述基于面部识别的痤疮判断方法、终端及存储介质，可以快速准确的对患者的情况做一个初步的判断并给出合适的治疗方案，进而提高患者就诊效率。

Description

基于面部识别的痤疮判断方法、终端及存储介质
本申请要求于2018年5月23日提交中国专利局，申请号为201810502333.2、发明名称为“基于面部识别的痤疮判断方法、终端及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及面部识别领域，尤其涉及一种基于面部识别的痤疮判断方法、终端及存储介质。
背景技术
痤疮是一种由多种因素引起的常见的慢性毛囊、皮脂腺炎症。在我国，人群中约有85％在12-24岁期间出现痤疮，其中仅有10％-30％的痤疮患者就诊于皮肤科，但国内仅有约2.2万名皮肤科医生，医生人数短缺加上资源分布不均，使许多痤疮患者在病发初期难以获得及时有效的治疗，最终导致症状恶化，甚至最终引发焦虑症、抑郁症等严重心理问题。
目前，痤疮诊治基本靠皮肤科医生的临床经验积累，但受专业医护人员资源的限制，无论是痤疮患者还是一些边缘地区接诊的医生都无法快速准确的获得一个对当前病症的诊断结果以及合适的治疗方案，进而耽误病人病情。
发明内容
有鉴于此，本申请提出一种基于面部识别的痤疮判断方法、终端及存储介质，通过实施上述方式，可以快速准确的对患者的情况做一个初步的判断，进而提高患者就诊效率。
首先，为实现上述目的，本申请提出一种终端，所述终端包括存储器、处理器，所述存储器上存储有可在所述处理器上运行的基于面部识别的痤疮 判断程序，所述基于面部识别的痤疮判断程序被所述处理器执行时实现如下步骤：获取训练后的深度学习模型；加载所述训练后的深度学习模型；采集面部图像；根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别；获取图像识别结果，并将所述图像识别结果上传至云端病理数据库，所述云端病理数据库用于根据所述图像识别结果进行病理检索；及接收所述云端病理数据库的病理检索结果并呈现给用户。
此外，为实现上述目的，本申请还提供一种基于面部识别的痤疮判断方法，所述基于面部识别的痤疮判断方法包括：获取训练后的深度学习模型；加载所述训练后的深度学习模型；采集面部图像；根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别；获取图像识别结果，并将所述图像识别结果上传至云端病理数据库，所述云端病理数据库用于根据所述图像识别结果进行病理检索；及接收所述云端病理数据库的病理检索结果并呈现给用户。
进一步地，为实现上述目的，本申请还提供一种存储介质，所述存储介质存储有基于面部识别的痤疮判断程序，所述基于面部识别的痤疮判断程序可被至少一个处理器执行，以使所述至少一个处理器执行如上所述的基于面部识别的痤疮判断方法的步骤。
附图说明
图1是实现本申请各个实施例的一种终端的硬件结构示意图；
图2是本申请一种终端的应用环境图；
图3是本申请基于面部识别的痤疮判断程序第一实施例的程序模块图；
图4是本申请基于面部识别的痤疮判断方法第一实施例的实施流程示意图；
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。
在后续的描述中，使用用于表示元件的诸如“模块”、“部件”或“单元”的后缀仅为了有利于本申请的说明，其本身没有特定的意义。因此，“模块”、“部件”或“单元”可以混合地使用。
终端可以以各种形式来实施。例如，本申请中描述的终端可以包括诸如手机、平板电脑、笔记本电脑、掌上电脑、个人数字助理(Personal Digital Assistant，PDA)、便捷式媒体播放器(Portable Media Player，PMP)、导航装置、可穿戴设备、智能手环、计步器等可具有图像获取功能的终端，以及诸如数字TV、台式计算机等可具有图像获取功能的固定终端。
后续描述中将以移动终端为例进行说明，本领域技术人员将理解的是，除了特别用于移动目的的元件之外，根据本申请的实施方式的构造也能够应用于固定类型的终端。
请参阅图1，其为实现本申请各个实施例的一种终端的硬件结构示意图，该终端100可以包括：RF(Radio Frequency，射频)单元101、WiFi模块102、音频输出单元103、A/V(音频/视频)输入单元104、传感器105、显示单元106、用户输入单元107、接口单元108、存储器109、处理器110、以及电源111等部件。本领域技术人员可以理解，图1中示出的终端装置结构并不构成对终端的限定，终端装置可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。其中存储器109可用于存储软件程序以及各种数据。存储器109可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需的应用程序(比如声音播放功能、图像播放功能等)等；存储数据区可存储根据手机的使用所创建的数据(比如音频数据、电话本等)等。此外，存储器109可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他易失性固态存储器件。
尽管图1未示出，终端100还可以包括蓝牙模块等，在此不再赘述。
图2是本申请一种终端的应用环境图。其基于图1所示的终端100，通过采集目标对象400的目标图像，并对采集的图像进行一定的初步处理后，导入训练好的深度学习模型进行计算，获得计算结果，并将计算结果上传至云端病理数据库500，所述云端病理数据库500用于根据所述图像识别结果进行病理检索；再接收所述云端病理数据库500的病理检索结果并呈现给用户。如此，可以利用深度学习，准确快速的识别症状，再结合病理数据库进行快速的病理检索，进而供患者或医生进行参考判断，提升了痤疮治疗效率。
基于上述终端100硬件结构、通信网络系统及上述应用环境，提出本申请方法各个实施例。
首先，本申请提出一种基于面部识别的痤疮判断程序300。
参阅图3所示，是本申请基于面部识别的痤疮判断程序300第一实施例的程序模块图。
本实施例中，所述的基于面部识别的痤疮判断程序300包括一系列的存储于上述终端100中存储器109上的计算机程序指令，当该计算机程序指令被处理器110执行时，可以实现本申请各实施例的基于面部识别的痤疮判断操作。在一些实施例中，基于该计算机程序指令各部分所实现的特定的操作，所述基于面部识别的痤疮判断程序300可以被划分为一个或多个模块。例如，在图3中，所述的基于面部识别的痤疮判断程序300可以被分割成训练模块301、加载模块302、采集模块303、识别模块304、检索模块305、输出模块306。其中：
所述训练模块301用于获取训练后的深度学习模型。
在本实施方式中，深度学习模型可以为VGG-16模型，也可以为inceptionV3模型，上述两种深度学习模型均为图像分类模型，为谷歌(Google)官网开放的深度学习模型代码。下载VGG-16模型或inceptionV3模型代码后，利用训练数据对上述VGG-16模型或inceptionV3模型进行训练，最终获得训练后的深度学习模型。
具体的，所述获取训练后的深度学习模型的步骤具体包括：
1)整理训练数据，所述训练数据为大量的痤疮病症图片。
在本实施方式中，痤疮可根据临床表现分为8类：粉刺性痤疮、丘疹性痤疮、脓包性痤疮、囊肿性痤疮、结节性痤疮、萎缩性痤疮、聚合性痤疮、恶病质性青春痘。而根据痤疮皮损性质和严重程度将痤疮分为4级：I级，总皮损数小于30个；II级，总皮损数31-50个；Ⅲ级，总皮损数50-100个,结节数小于3个；Ⅳ级，总皮损数大于100个，结节/囊肿数大于3个。故基于此，对用户上传的所有训练数据分类为4*8＝32类，并以一定格式保存，即将同一类数据存放于同一个文件夹，文件名即为其类名。
2)获取用户上传的规定格式的训练数据。
3)验证获取的训练数据的格式，即判断是否符合步骤1)中确定的数据格式。
4)利用迁移学习方式将所述训练数据载入深度学习模型进行训练以获取训练后的深度学习模型，即通过深度学习模型对载入的训练数据进行预订次数的迭代计算，进而得到训练后的深度学习模型。
加载模块302用于加载所述训练后的深度学习模型。
在本实施方式中，上述终端中包括tensorflow模块，其中所述tensorflow是谷歌(Google)基于DistBelief进行研发的第二代人工智能学习系统，其可以支持深度神经学习网络进行流图的快速计算。具体的，上述tensorflow模块在通过加载完成上述训练后的深度学习模型之后，还对所述深度学习模型进行初始化，进而为后续深度学习模型的计算做好准备。
采集模块303用于采集面部图像。
在本实施方式中，采集面部图像主要通过调用终端的图像获取装置的API接口，再通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集的面部图像。其中，所述图像获取装置优选为摄像头。
更进一步地，采集模块303在采集面部图像的的过程中：首先通过终端显示界面提示用户选择上传照片或拍摄照片；若用户选择上传照片，提示用户上传正向、左向、右向面部照片各一张作为所述面部图像；若用户选择拍摄照片，提示用户将面部暴露在自然光源下，并拍摄正向、左向、右向面部照片各一张作为所述面部图像。
更进一步地，采集模块303在采集完所述面部图像之后，还可以对所述面部图像进行预处理，即对采集的面部图像进行预设尺寸的裁剪，并对裁剪后的面部图像进行压缩。因为一般采集的面部图像都是1M以上的大小，图片过大对计算过程中数据传输效率存在影响。故通过预先的裁剪和压缩，提高上述深度学习模型计算的效率。
识别模块304用于根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别。
在本实施方式中，识别模块304将采集的所述面部图像发送到上述tensorflow模块中，所述tensorflow模块将所述面部图像在所述训练后的深度学习模型中计算。具体而言，任何图像本质对于计算机而言都是一定大小的数字阵列，比如讲所述面部图像裁剪为18*18像素尺寸的大小，输入到训练后的深度学习模型中(此时深度学习模型可以设置为324个输入节点)，所述深度学习模型针对4*8＝32个输出均会反复迭代计算出一个结果。
检索模块305用于获取图像识别结果，并将所述图像识别结果上传至云端病理数据库500，所述云端病理数据库500用于根据所述图像识别结果进行病理检索。
在本实施方式中，上述图像识别结果为32个输出结果中数值最大的输出端口代表的类，再将该结果上传至上述云端病理数据500中，上述云端病理数据库500存放着各种病症对应的症状、治疗方法、注意事项等病理。比如I级粉刺性痤疮对应的症状图片、如何治疗、治疗期和康复期的注意事项等等。
输出模块306用于接收所述云端病理数据库的病理检索结果并呈现给用 户。
在本实施方式中，所述输出模块306可以通过终端100的显示单元106进行匹配结果的输出，进而为用户提供一个直观的病情初步诊断。
通过上述程序模块301-306，本申请所提出的基于面部识别的痤疮判断程序300，通过对采集的面部图像进行识别后，快速的获取相应的诊断结果及治疗信息，进而为患者提供更高效的治疗方案。
此外，本申请还提出一种基于面部识别的痤疮判断方法。
参阅图4所示，是本申请基于面部识别的痤疮判断方法第一实施例的实施流程示意图。
步骤S401，获取训练后的深度学习模型。
在本实施方式中，深度学习模型可以为VGG-16模型，也可以为inceptionV3模型，上述两种深度学习模型均为图像分类模型，为谷歌(Google)官网开放的深度学习模型代码。下载VGG-16模型或inceptionV3模型代码后，利用训练数据对上述VGG-16模型或inceptionV3模型进行训练，最终获得训练后的深度学习模型。
具体的，所述获取训练后的深度学习模型的步骤具体包括：
5)整理训练数据，所述训练数据为大量的痤疮病症图片。
在本实施方式中，痤疮可根据临床表现分为8类：粉刺性痤疮、丘疹性痤疮、脓包性痤疮、囊肿性痤疮、结节性痤疮、萎缩性痤疮、聚合性痤疮、恶病质性青春痘。而根据痤疮皮损性质和严重程度将痤疮分为4级：I级，总皮损数小于30个；II级，总皮损数31-50个；Ⅲ级，总皮损数50-100个,结节数小于3个；Ⅳ级，总皮损数大于100个，结节/囊肿数大于3个。故基于此，对用户上传的所有训练数据分类为4*8＝32类，并以一定格式保存，即将同一类数据存放于同一个文件夹，文件名即为其类名。
6)获取用户上传的规定格式的训练数据。
7)验证获取的训练数据的格式，即判断是否符合步骤1)中确定的数据格式。
8)利用迁移学习方式将所述训练数据载入深度学习模型进行训练以获取训练后的深度学习模型，即通过深度学习模型对载入的训练数据进行预订次数的迭代计算，进而得到训练后的深度学习模型。
步骤S402，加载所述训练后的深度学习模型。
在本实施方式中，上述终端中包括tensorflow模块，其中所述tensorflow是谷歌(Google)基于DistBelief进行研发的第二代人工智能学习系统，其可以支持深度神经学习网络进行流图的快速计算。具体的，上述tensorflow模块在通过加载完成上述训练后的深度学习模型之后，还对所述深度学习模型进行初始化，进而为后续深度学习模型的计算做好准备。
步骤S403，采集面部图像。
在本实施方式中，采集面部图像主要通过调用终端的图像获取装置的API接口，再通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集的面部图像。其中，所述图像获取装置优选为摄像头。
更进一步地，采集模块303在采集面部图像的的过程中：首先通过终端显示界面提示用户选择上传照片或拍摄照片；若用户选择上传照片，提示用户上传正向、左向、右向面部照片各一张作为所述面部图像；若用户选择拍摄照片，提示用户将面部暴露在自然光源下，并拍摄正向、左向、右向面部照片各一张作为所述面部图像。
更进一步地，采集模块303在采集完所述面部图像之后，还可以对所述面部图像进行预处理，即对采集的面部图像进行预设尺寸的裁剪，并对裁剪后的面部图像进行压缩。因为一般采集的面部图像都是1M以上的大小，图片过大对计算过程中数据传输效率存在影响。故通过预先的裁剪和压缩，提高上述深度学习模型计算的效率。
步骤S404，根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别。
在本实施方式中，识别模块304将采集的所述面部图像发送到上述tensorflow模块中，所述tensorflow模块将所述面部图像在所述训练后的深度学习模型中计算。具体而言，任何图像本质对于计算机而言都是一定大小的数字阵列，比如讲所述面部图像裁剪为18*18像素尺寸的大小，输入到训练后的深度学习模型中(此时深度学习模型可以设置为324个输入节点)，所述深度学习模型针对4*8＝32个输出均会反复迭代计算出一个结果。
步骤S405，获取图像识别结果，并将所述图像识别结果上传至云端病理数据库500，所述云端病理数据库500用于根据所述图像识别结果进行病理检索。
在本实施方式中，上述图像识别结果为32个输出结果中数值最大的输出端口代表的类，再将该结果上传至上述云端病理数据500中，上述云端病理数据库500存放着各种病症对应的症状、治疗方法、注意事项等病理。比如I级粉刺性痤疮对应的症状图片、如何治疗、治疗期和康复期的注意事项等等。
在步骤S406，接收所述云端病理数据库的病理检索结果并呈现给用户。
在本实施方式中，通过终端100的显示单元106进行匹配结果的输出，进而为用户提供一个直观的病情初步诊断。
通过执行上述步骤S401-S406，本申请所提出的基于面部识别的痤疮判断方法，通过对采集的面部图像进行识别后，快速的获取相应的诊断结果及治疗信息，进而为患者提供更高效的治疗方案。
本申请还提供了另一种实施方式，即提供一种存储介质，所述存储介质存储有基于面部识别的痤疮判断程序，所述基于面部识别的痤疮判断程序可被至少一个处理器执行，以使所述至少一个处理器执行如上所述的基于面部识别的痤疮判断方法的步骤。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种基于面部识别的痤疮判断方法，应用于终端，其特征在于，所述方法包括步骤：
    获取训练后的深度学习模型；
    加载所述训练后的深度学习模型；
    采集面部图像；
    根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别；
    获取图像识别结果，并将所述图像识别结果上传至云端病理数据库，所述云端病理数据库用于根据所述图像识别结果进行病理检索；及
    接收所述云端病理数据库的病理检索结果并呈现给用户。
  2. 如权利要求1所述的基于面部识别的痤疮判断方法，其特征在于，所述获取训练后的深度学习模型的步骤具体包括：
    获取用户上传的规定格式的训练数据；
    利用迁移学习方式将所述训练数据载入深度学习模型进行训练以获取训练后的深度学习模型。
  3. 如权利要求1所述的基于面部识别的痤疮判断方法，其特征在于，所述终端包括tensorflow模块，所述加载所述训练后的深度学习模型的步骤具体包括：
    通过tensorflow模块加载所述训练后的深度学习模型并进行初始化，所述所述训练后的深度学习模型为训练后的VGG-16模型或inceptionV3模型。
  4. 如权利要求3所述的基于面部识别的痤疮判断方法，其特征在于，所述根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别的步骤具体包括：
    将采集的所述面部图像发送到所述tensorflow模块中，所述tensorflow模块将所述面部图像在所述训练后的深度学习模型中计算。
  5. 如权利要求1所述的基于面部识别的痤疮判断方法，其特征在于，所 述采集面部图像的步骤包括：
    调用终端的图像获取装置的API接口；
    通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集的面部图像。
  6. 如权利要求2-4任一项所述的基于面部识别的痤疮判断方法，其特征在于，所述采集面部图像的步骤包括：
    调用终端的图像获取装置的API接口；
    通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集的面部图像。
  7. 如权利要求6所述的基于面部识别的痤疮判断方法，其特征在于，所述采集面部图像的步骤具体还包括：
    对采集的所述面部图像进行预设尺寸的裁剪；
    对裁剪后的所述面部图像进行压缩。
  8. 一种终端，其特征在于，所述终端包括存储器、处理器，所述存储器上存储有可在所述处理器上运行的基于面部识别的痤疮判断程序，所述基于面部识别的痤疮判断程序被所述处理器执行时实现如下步骤：
    获取训练后的深度学习模型；
    加载所述训练后的深度学习模型；
    采集面部图像；
    根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别；
    获取图像识别结果，并将所述图像识别结果上传至云端病理数据库，所述云端病理数据库用于根据所述图像识别结果进行病理检索；及
    接收所述云端病理数据库的病理检索结果并呈现给用户。
  9. 如权利要求8所述的终端，其特征在于，所述处理器处理所述获取训练后的深度学习模型的步骤之前还包括：
    获取用户上传的规定格式的训练数据；
    利用迁移学习方式将所述训练数据载入深度学习模型进行训练以获取训练后的深度学习模型。
  10. 如权利要求8所述的终端，其特征在于，所述终端包括tensorflow模块，所述加载所述训练后的深度学习模型的步骤具体包括：
    通过tensorflow模块加载所述训练后的深度学习模型并进行初始化，所述所述训练后的深度学习模型为训练后的VGG-16模型或inceptionV3模型。
  11. 如权利要求10所述的终端，其特征在于，所述根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别的步骤具体包括：
    将采集的所述面部图像发送到所述tensorflow模块中，所述tensorflow模块将所述面部图像在所述训练后的深度学习模型中计算。
  12. 如权利要求8所述的终端，其特征在于，所述采集面部图像的步骤包括：
    调用终端的图像获取装置的API接口；
    通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集的面部图像。
  13. 如权利要求9-11任一项所述的终端，其特征在于，所述采集面部图像的步骤包括：
    调用终端的图像获取装置的API接口；
    通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集的面部图像。
  14. 如权利要求13所述的终端，其特征在于，所述采集面部图像的步骤具体还包括：
    对采集的所述面部图像进行预设尺寸的裁剪；
    对裁剪后的所述面部图像进行压缩。
  15. 一种计算机可读存储介质，所述计算机可读存储介质存储有基于面部识别的痤疮判断程序，所述基于面部识别的痤疮判断程序可被至少一个处理 器执行，以使所述至少一个处理器执行如下步骤：
    获取训练后的深度学习模型；
    加载所述训练后的深度学习模型；
    采集面部图像；
    根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别；
    获取图像识别结果，并将所述图像识别结果上传至云端病理数据库，所述云端病理数据库用于根据所述图像识别结果进行病理检索；及
    接收所述云端病理数据库的病理检索结果并呈现给用户。
  16. 如权利要求15所述的计算机可读存储介质，其特征在于，所述处理器处理所述获取训练后的深度学习模型的步骤之前还包括：
    获取用户上传的规定格式的训练数据；
    利用迁移学习方式将所述训练数据载入深度学习模型进行训练以获取训练后的深度学习模型。
  17. 如权利要求15所述的计算机可读存储介质，其特征在于，所述终端包括tensorflow模块，所述加载所述训练后的深度学习模型的步骤具体包括：
    通过tensorflow模块加载所述训练后的深度学习模型并进行初始化，所述所述训练后的深度学习模型为训练后的VGG-16模型或inceptionV3模型。
  18. 如权利要求17所述的计算机可读存储介质，其特征在于，所述根据采集的所述面部图像和所述训练后的深度学习模型进行图像识别的步骤具体包括：
    将采集的所述面部图像发送到所述tensorflow模块中，所述tensorflow模块将所述面部图像在所述训练后的深度学习模型中计算。
  19. 如权利要求15-18任一项所述的计算机可读存储介质，其特征在于，所述采集面部图像的步骤包括：
    调用终端的图像获取装置的API接口；
    通过所述API接口开启所述图像获取装置并接收所述图像获取装置采集 的面部图像。
  20. 如权利要求19所述的计算机可读存储介质，其特征在于，所述采集面部图像的步骤具体还包括：
    对采集的所述面部图像进行预设尺寸的裁剪；
    对裁剪后的所述面部图像进行压缩。
PCT/CN2018/102201 2018-05-23 2018-08-24 基于面部识别的痤疮判断方法、终端及存储介质 Ceased WO2019223148A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810502333.2A CN108876766A (zh) 2018-05-23 2018-05-23 基于面部识别的痤疮判断方法、终端及存储介质
CN201810502333.2 2018-05-23

Publications (1)

Publication Number Publication Date
WO2019223148A1 true WO2019223148A1 (zh) 2019-11-28

Family

ID=64333838

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/102201 Ceased WO2019223148A1 (zh) 2018-05-23 2018-08-24 基于面部识别的痤疮判断方法、终端及存储介质

Country Status (2)

Country Link
CN (1) CN108876766A (zh)
WO (1) WO2019223148A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210287797A1 (en) * 2020-03-11 2021-09-16 Memorial Sloan Kettering Cancer Center Parameter selection model using image analysis
CN117333487A (zh) * 2023-12-01 2024-01-02 深圳市宗匠科技有限公司 一种痘痘分级方法、装置、设备及存储介质

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111382294A (zh) * 2018-12-11 2020-07-07 上海维域信息科技有限公司 一种基于人工智能图像识别的中医辅助判断方法
CN110473199B (zh) * 2019-08-21 2022-09-27 广州纳丽生物科技有限公司 基于深度学习实例分割的色斑痤疮检测与健康评价方法
CN112949667A (zh) * 2019-12-09 2021-06-11 北京金山云网络技术有限公司 图像识别方法、系统、电子设备及存储介质
KR102271558B1 (ko) * 2020-08-03 2021-07-01 아람휴비스 주식회사 빅데이터를 활용한 인공지능 두피영상 진단 분석 시스템 및 이를 이용한 제품 추천 시스템
CN112767383B (zh) * 2021-01-29 2024-02-27 深圳艾摩米智能科技有限公司 人脸痘痘定位识别方法
CN113128375B (zh) * 2021-04-02 2024-05-10 西安融智芙科技有限责任公司 图像识别方法及电子设备、计算机可读存储介质
CN113159227A (zh) * 2021-05-18 2021-07-23 中国医学科学院皮肤病医院(中国医学科学院皮肤病研究所) 一种基于神经网络的痤疮图像识别方法、系统和装置
CN114678124A (zh) * 2022-03-09 2022-06-28 武汉大学 一种压疮等级检测方法、装置、电子设备及可读存储介质
CN115509147B (zh) * 2022-08-25 2025-05-16 丝路视觉科技股份有限公司 设备管控方法、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030065552A1 (en) * 2001-10-01 2003-04-03 Gilles Rubinstenn Interactive beauty analysis
CN104299011A (zh) * 2014-10-13 2015-01-21 吴亮 一种基于人脸图像识别的肤质与皮肤问题识别检测方法
CN107194158A (zh) * 2017-05-04 2017-09-22 深圳美佳基因科技有限公司 一种基于图像识别的疾病辅助诊断方法
CN107491652A (zh) * 2017-08-30 2017-12-19 广州云从信息科技有限公司 一种基于人脸识别的人脸异常特征数据分析方法
CN107945173A (zh) * 2017-12-11 2018-04-20 深圳市宜远智能科技有限公司 一种基于深度学习的皮肤疾病检测方法以及系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030065552A1 (en) * 2001-10-01 2003-04-03 Gilles Rubinstenn Interactive beauty analysis
CN104299011A (zh) * 2014-10-13 2015-01-21 吴亮 一种基于人脸图像识别的肤质与皮肤问题识别检测方法
CN107194158A (zh) * 2017-05-04 2017-09-22 深圳美佳基因科技有限公司 一种基于图像识别的疾病辅助诊断方法
CN107491652A (zh) * 2017-08-30 2017-12-19 广州云从信息科技有限公司 一种基于人脸识别的人脸异常特征数据分析方法
CN107945173A (zh) * 2017-12-11 2018-04-20 深圳市宜远智能科技有限公司 一种基于深度学习的皮肤疾病检测方法以及系统

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210287797A1 (en) * 2020-03-11 2021-09-16 Memorial Sloan Kettering Cancer Center Parameter selection model using image analysis
US11887732B2 (en) * 2020-03-11 2024-01-30 Memorial Sloan Kettering Cancer Center Parameter selection model using image analysis
CN117333487A (zh) * 2023-12-01 2024-01-02 深圳市宗匠科技有限公司 一种痘痘分级方法、装置、设备及存储介质
CN117333487B (zh) * 2023-12-01 2024-03-29 深圳市宗匠科技有限公司 一种痘痘分级方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN108876766A (zh) 2018-11-23

Similar Documents

Publication Publication Date Title
WO2019223148A1 (zh) 基于面部识别的痤疮判断方法、终端及存储介质
CN108230296B (zh) 图像特征的识别方法和装置、存储介质、电子装置
CN111680678B (zh) 目标区域识别方法、装置、设备及可读存储介质
CN107680684B (zh) 用于获取信息的方法及装置
US12141983B2 (en) Image processing method and apparatus for segmenting objects
CN113241175B (zh) 基于边缘计算的帕金森症辅助诊断系统和方法
CN107729929A (zh) 用于获取信息的方法及装置
WO2023173646A1 (zh) 一种表情识别方法及装置
CN107526769A (zh) 基于人工智能的照片处理方法及装置、设备及可读介质
CN111209916A (zh) 病灶识别方法及系统、识别设备
CN104408678A (zh) 一种供个人使用的电子病历系统
CN110610125A (zh) 基于神经网络的牛脸识别方法、装置、设备及存储介质
CN118626615A (zh) 自动问答方法、装置、电子设备及计算机存储介质
KR102594093B1 (ko) 딥러닝 기술을 활용한 피부과 시술 추천 시스템 및 방법
CN111985197B (zh) 一种基于医疗信息的模板生成方法
CN110809037B (zh) 一种基于深度多元特征的物联网皮肤镜系统
CN116597176A (zh) 图像去重方法、装置、设备及存储介质
CN114464328A (zh) 试验信息的检索方法及装置、临床试验的推荐方法、终端
Oliveira et al. A bag-of-tasks approach to speed up the lung nodules retrieval in the bigdata age
CN118173287B (zh) 一种医疗视频在线诊疗方法及系统
CN119361129A (zh) 一种基于det评估量表的造口周围皮肤ai评估方法
CN108021648B (zh) 一种搜题的方法、装置及智能终端
WO2024251271A1 (zh) 图像分类方法和系统
CN114328969A (zh) 基于知识图谱的健康管理方法、装置、设备和介质
CN110570943B (zh) 智能推荐mdt入组的方法及装置、电子设备、存储介质

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: 18919525

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 25.03.2021)

122 Ep: pct application non-entry in european phase

Ref document number: 18919525

Country of ref document: EP

Kind code of ref document: A1