WO2019072249A1 - 利用皮肤图像分析皮肤水分的方法及装置 - Google Patents

利用皮肤图像分析皮肤水分的方法及装置 Download PDF

Info

Publication number
WO2019072249A1
WO2019072249A1 PCT/CN2018/110130 CN2018110130W WO2019072249A1 WO 2019072249 A1 WO2019072249 A1 WO 2019072249A1 CN 2018110130 W CN2018110130 W CN 2018110130W WO 2019072249 A1 WO2019072249 A1 WO 2019072249A1
Authority
WO
WIPO (PCT)
Prior art keywords
skin
image
brightness
value
roughness
Prior art date
Application number
PCT/CN2018/110130
Other languages
English (en)
French (fr)
Inventor
邱显荣
辛琳霖
周利民
Original Assignee
精诚工坊电子集成技术（北京）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 精诚工坊电子集成技术（北京）有限公司 filed Critical 精诚工坊电子集成技术（北京）有限公司
Publication of WO2019072249A1 publication Critical patent/WO2019072249A1/zh
Priority to US16/846,326 priority Critical patent/US11490854B2/en

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/44—Detecting, measuring or recording for evaluating the integumentary system, e.g. skin, hair or nails
    • A61B5/441—Skin evaluation, e.g. for skin disorder diagnosis
    • A61B5/442—Evaluating skin mechanical properties, e.g. elasticity, hardness, texture, wrinkle assessment
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
    • A61B5/0062—Arrangements for scanning
    • A61B5/0064—Body surface scanning
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/44—Detecting, measuring or recording for evaluating the integumentary system, e.g. skin, hair or nails
    • A61B5/441—Skin evaluation, e.g. for skin disorder diagnosis
    • A61B5/443—Evaluating skin constituents, e.g. elastin, melanin, water
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • G06T5/30—Erosion or dilatation, e.g. thinning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/40—Analysis of texture
    • G06T7/41—Analysis of texture based on statistical description of texture
    • G06T7/44—Analysis of texture based on statistical description of texture using image operators, e.g. filters, edge density metrics or local histograms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/54—Extraction of image or video features relating to texture
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/60—Extraction of image or video features relating to illumination properties, e.g. using a reflectance or lighting model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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 invention relates to a method for analyzing skin moisture by using skin images, and also relates to a device for analyzing skin moisture by using skin images, and belongs to the technical field of skin detection.
  • Skin detectors that use skin testers to measure skin moisture have been widely used and have received great attention in skin care and medical applications.
  • the Chinese invention patent No. ZL 201310645552.3 discloses a skin moisture detector. It uses multiple electrodes to attach to multiple parts of the human body, and sends a small current signal to detect the skin capacitance distribution, thereby obtaining skin moisture information of the corresponding area.
  • Chinese Patent Application No. 201510403775.8 discloses a skin parameter measuring circuit and method. It also uses the electrode to contact the skin, the skin resistance is measured by the moisture detection module, and the skin moisture is detected according to the relationship between the skin resistance and the skin moisture.
  • the parameters such as the capacitance or the capacitive reactance of the skin are generally obtained by using techniques such as sensors and electrodes, and the corresponding relationship between the capacitance or the capacitive reactance and the skin moisture is utilized to obtain the moisture state of the skin.
  • the primary technical problem to be solved by the present invention is to provide a method for analyzing skin moisture using a skin image.
  • Another technical problem to be solved by the present invention is to provide an apparatus for analyzing skin moisture using a skin image.
  • a method of analyzing skin moisture using a skin image comprising the steps of:
  • Skin image analysis is performed on the obtained skin image, and skin moisture is obtained based on the skin characteristics.
  • the skin characteristics include gloss and smoothness.
  • the skin moisture is obtained based on the skin characteristics, and the skin is judged to be oily skin when the gloss is shiny; when the gloss is lack of gloss, the skin is judged to be dry skin.
  • the skin is judged to be oily skin; if the roughness is rough, the skin is judged to be dry skin; If the roughness is smooth and fine, the skin is judged to be neutral skin.
  • the number of pixels whose brightness is greater than the mean value of the brightness is divided by the total number of pixels of the skin image, and the percentage of pixels of the high brightness is obtained;
  • the skin image with the high-brightness pixel point percentage greater than M1 is determined to be shiny; the high-brightness pixel percentage is less than or equal to M1 and greater than M2.
  • the skin image is determined to be glossy; the skin image with a high brightness pixel percentage less than or equal to M2 is determined to be lack of gloss.
  • the smoothness is obtained based on the skin roughness and the average pore area.
  • the calculation of the skin roughness comprises the following steps:
  • Step 32 Improve the contrast and brightness of the skin image by the SSR algorithm
  • Step 33 Calculate the gradation difference between the pixels in the horizontal direction, the diagonal direction, the vertical direction, and the diagonal direction by using the gray level co-occurrence matrix method, obtain the contrast, and find the contrast in the four directions;
  • Step 34 Using the roughness threshold value for presetting the classification, the roughness classification is performed according to the contrast variance value, and the texture is rough, rough, and delicate.
  • the calculation of the pore area comprises the following steps:
  • Step 36 Perform color space conversion on the image processed by the SSR algorithm, and convert the image into an HSV color space.
  • Step 37 Calculate the saturation value S in the HSV color space to binarize the image
  • Step 38 Calculate the average pore area and pore area
  • the skin image is determined to be smooth and delicate; when the roughness value is rough and the average pore area is normal, the judgment is performed.
  • the image of the skin is relatively rough; when the roughness value is rough and the average area of the pores is coarse, the skin image is judged to be coarse and coarse.
  • an apparatus for analyzing skin moisture using a skin image comprising:
  • a processor that controls the lens module, configured to execute an instruction to perform skin feature analysis on the obtained skin image, and obtain skin moisture based on the skin feature;
  • a memory coupled to the processor for storing executable instructions of the processor.
  • the method and device for analyzing skin moisture using skin images provided by the present invention significantly reduce the implementation cost compared with the prior art, and can realize rapid detection. Further, in the case where the skin tester is not used for the condition, the skin condition of the skin can be directly obtained by the skin image, and the use range of the skin moisture analysis can be effectively expanded.
  • FIG. 1 is a flow chart of a method for analyzing skin moisture using a skin image according to the present invention
  • Figure 2 is a flow chart of the gloss analysis step of Figure 1;
  • FIG. 3 is a schematic diagram showing the relationship between a reflectance image and an illumination image in an SSR algorithm
  • Figure 5 is a schematic diagram of the gloss output value of Figure 2;
  • Figure 6 is a schematic diagram of the smoothness output value of Figure 1;
  • FIG. 7 is a schematic diagram showing a black and white boundary of a skin image after binarization
  • Figure 8 is a flow chart showing the qualitative determination steps of the skin moisture condition of Figure 1;
  • Figure 9 is a schematic view of a device for analyzing skin moisture using a skin image according to the present invention.
  • the method specifically includes the following steps:
  • Step 1 Get a skin image
  • a skin image such as a skin image taken by a skin tester. It can be a skin image of the whole face, a skin image of a part of the face, or a skin image of other parts of the human body.
  • low-pass filtering, erosion, and expansion can also be performed to remove unnecessary noise and pre-process the skin image.
  • Step 2 Calculate the gloss value
  • the present invention calculates the gloss value of the skin image based on the pre-processed skin image, and obtains the classification and result of the gloss. Gloss is divided into shiny, shiny, and lacking luster.
  • the method for calculating the gloss value may be various. The calculation method shown in FIG. 2 is adopted in the embodiment of the present invention, and will be described in detail later with reference to FIGS. 2 to 5.
  • Step 21 Read the skin image
  • the acquired skin image is converted into a HSL color space by color space conversion, where L represents brightness.
  • Step 22 Improve the contrast and brightness of the image by the SSR algorithm.
  • the SSR algorithm can convert the image color into a color value similar to that recognized by the human eye.
  • the SSR (Scalable Sample Rate) algorithm is an image enhancement algorithm based on Retinex theory. Based on Retinex theory, an image can be divided into an incident component and a reflection component. For a reflection component, a Gaussian function is used to convolve the component, and then the two components are added to calculate a final result.
  • Step 23 Image preprocessing is performed by low-pass filtering, and the illumination image L(x, y) is obtained from the preprocessed image S(x, y), and the reflectance image R(x, y) is calculated by the following formula.
  • Step 24 The image processed by the SSR algorithm is analyzed, and the distribution of the luminance value L is analyzed to obtain a luminance average.
  • the luminance average (which may be an arithmetic mean or a geometric mean, etc.) may be obtained by simple calculation.
  • Step 25 Count the percentage of pixels that are greater than the mean of the brightness.
  • the number of pixels (high-brightness pixels) whose luminance value L is larger than the luminance average value is counted, and then the number of high-brightness pixel points is divided by the total number of pixel points in the image as a percentage of high-brightness pixel points.
  • the skin image with the high-brightness pixel point percentage greater than M1 is determined to be shiny; the high-brightness pixel percentage is less than or equal to M1 and greater than M2.
  • the skin image is determined to be glossy; the skin image with a high brightness pixel percentage less than or equal to M2 is determined to be lack of gloss.
  • M1 and M2 are obtained based on the skin image big data statistics.
  • the brightness pixel percentage and the high brightness pixel percentage are summed to obtain the sum of the percentages; according to the sum of the percentages, compared with the pixel point percentage thresholds M1 and M2, if the sum of the percentages is greater than M1, it is judged that the shine is bright; If the sum of the percentages is less than or equal to M1 and greater than M2, it is judged to be glossy; if the sum of the percentages is less than or equal to M2, it is determined to be lack of gloss.
  • Step 3 Calculate the smoothness value
  • the present invention is based on a pre-processed skin image, and the smoothness value is obtained by calculating the skin roughness and the pore area in the skin image, and the calculation result of the smoothness is shown in Fig. 6.
  • the smoothness value calculation in the present invention is obtained based on the pore value and the texture value.
  • the calculation of the pore value and the texture value can be carried out by various calculation methods commonly used in the prior art. E.g:
  • the method of calculating the mean value of the deviation absolute value as the roughness feature value of the skin image using the color space pixel value of the skin image disclosed in the prior patent application of the applicant No. 201710337597.2 converts a face image from an RGB color space to a YCbCr chromaticity space, and extracts a luminance component, a blue chrominance component, and a red chrominance component map. Get skin color similarity and roughness.
  • the patent application discloses a skin pore identification method based on image analysis, which calculates a luminance difference matrix of a pixel region; obtains a new image by fusing the original image and the luminance difference matrix, and performs cluster analysis on the new image; Image; count the number of skin pores and calculate the average pixel area of the skin pores.
  • the present invention can adopt the above calculation method of texture value and pore value, and can also calculate the roughness and the number of pores by the following calculation methods. Finally, the roughness is combined with the mean number of pores to divide the smoothness of the skin into smooth, fine, rough, coarse pores and coarse texture.
  • the steps to calculate the roughness are as follows:
  • Step 31 Get a skin image
  • Step 32 Enhance the contrast and brightness of the skin image by the SSR algorithm.
  • Step 33 Calculate the relative directions of 0 degrees (horizontal direction), 45 degrees (diagonal direction), 90 degrees (vertical direction), and 135 degrees (opposite direction) by GLCM (Grayscale Co-occurrence Matrix Method) algorithm. The difference in gradation between the two is obtained, and the contrast is obtained, and the contrast in the four directions is determined as a variance.
  • GLCM Gramscale Co-occurrence Matrix Method
  • the size of the main diagonal element value is used to judge the texture direction; using pixel values Discreteness, to reflect the thickness of the texture, the pixel value far from the main diagonal is high, that is, the dispersion is large, indicating that the gray scale difference of adjacent pixels is high, indicating that the texture perpendicular to the direction on the image is fine.
  • the GLCM algorithm can extract 14 texture features, where contrast (inertia moment) can reflect image sharpness and texture strength.
  • contrast inertia moment
  • the larger the contrast value the deeper the groove of the texture and the easier it is to be observed.
  • the variance will be larger. Textures tend to exhibit a single directionality, so the more complex the texture in contrast, the greater the difference in contrast across the four directions. Conversely, if the skin is smooth, the closer the contrast in the four directions, the smaller the variance will be.
  • Step 34 Perform roughness classification based on the contrast variance value.
  • the roughness threshold for classification is set in advance, and the roughness category is divided according to the threshold.
  • the threshold value selection for roughness is based on a large amount of skin data statistics and is stored in advance in the memory of the skin tester.
  • two roughness threshold values a first roughness threshold value and a second roughness threshold value (greater than the first roughness threshold value) are set. If the contrast variance value is less than the first roughness threshold value, it is determined to be smooth; if the contrast variance value is greater than the second roughness threshold value, it is determined that the texture is rough; if the contrast variance value is greater than or equal to the first roughness The threshold value is less than or equal to the second roughness threshold value, and is judged to be rough.
  • the step of calculating the mean pore area based on the skin image processed by the SSR algorithm includes:
  • Step 35 Obtain a skin image and perform SSR algorithm processing
  • step 35 has the same effect as step 31 and step 32 of calculating the roughness
  • the output of step 32 can be used to calculate the pore area directly in the actual calculation process when calculating the roughness.
  • Step 36 Perform color space conversion on the image processed by the SSR algorithm, and convert the image into an HSV color space.
  • Step 37 Calculate the saturation value S in the HSV color space.
  • the maximum inter-class variance method (also known as the Otsu method, referred to as OTSU) is used to calculate the binarization threshold value T.
  • the image is binarized by using the threshold value, and is divided into two parts: a high saturation zone and a low saturation zone.
  • the pores are considered to be dark parts relative to the skin and are low saturation areas; the skin is a highly saturated area. The larger the pores, the more obvious the difference in color saturation.
  • the main function of the largest inter-class variance is to find the threshold between the background and the target. Find the threshold by the amount of grayscale distribution.
  • the pores are the target and the skin is the background.
  • the larger the variance between the two regions the greater the difference between the two regions that make up the image.
  • the difference between the two regions will be smaller. Therefore, the maximum inter-class variance method is used to obtain the inter-class variance.
  • the segmentation threshold with the largest variance means that the probability of misclassifying the two regions is the smallest, and thus the results of the calculation of the pores obtained by the present invention are relatively accurate.
  • Step 38 Calculate the average pore area and pore area
  • each black and white border in the image is marked as a circle (as shown in Figure 7) as a pore.
  • Calculate the area of all circles and the number of circles, and calculate the mean area / number of circles as the mean area of the pores.
  • the pore area threshold value is pre-stored in the memory of the skin tester, and the average pore area is divided into pores and pores.
  • the skin image was divided as shown in Table 1, to obtain a smooth value.
  • Table 1 Table of smooth values based on the mean roughness and pore area
  • Step 4 Calculate skin moisture status by combining smoothness value and gloss value
  • the skin moisture condition obtained by the present invention is calculated based on the smoothness value and the gloss value of the image, rather than being directly detected by the sensor as in the prior art.
  • the qualitative determination steps for skin moisture conditions are as follows:
  • Step 41 Determine whether the gloss output result is shine based on the gloss value. If the gloss is shine, it can be concluded that the skin moisture results in oily skin; if it is not shine, go to the next step;
  • Step 42 judging whether the gloss output result lacks luster, if the gloss is lack of luster, it can be concluded that the skin moisture result is dry skin; if it is not lack of luster, that is, the skin is shiny, then the next step is entered;
  • Step 43 Determine whether the smoothness value is smooth and delicate. If it is smooth and delicate, it means that in this case, the gloss of the skin is shiny, and the smoothness is smooth and fine, then it can be concluded that the skin moisture results in neutral skin; if the smoothness is not smooth and fine, then proceed to the next step. ;
  • Step 44 Determine whether the gloss value is close to the shine value.
  • the approaching the oil value means that, when calculating the gloss value, the percentage of the pixel points in the skin image that is greater than the highlight threshold A is smaller than but close to M1, for example, the pixel in the skin image that is greater than the highlight threshold A.
  • the dot percentage is 85% to 99% of M1.
  • Step 45 If the gloss value is close to the shine value, it is judged to be oily skin; otherwise it is judged to be dry skin. In other words, with this step, the rough-skinned or rough-skinned rough skin is classified into oily skin or dry skin depending on whether the glossiness is close to the oil-light value M1.
  • the skin moisture condition can be obtained by a look-up table using the following table.
  • the invention utilizes skin image analysis to obtain skin moisture condition, on the one hand avoids dependence on the skin detector; on the other hand, the accuracy of the analysis result has been experimentally verified due to comprehensive consideration of multiple dimensions of roughness, pores and gloss.
  • the present invention also provides a device (i.e., a skin tester) that utilizes a skin image to obtain skin moisture.
  • the apparatus includes a lens module 10 having a large imaging area, a processor 20, and a memory 30 coupled to the processor 20.
  • the lens module 10 can realize a large shooting area;
  • the processor 20 is configured to execute a program;
  • the memory 30 is configured to store program instructions capable of executing various method steps when running on the processor 20, and for the processor to run the program
  • the threshold value and water condition judgment table for example, Table 2 that are called at the time.
  • the photographing area of the lens module of the present invention needs to be greater than or equal to 1.5 x 1.5 cm 2 so that sufficient skin characteristics can be obtained.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Public Health (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Biomedical Technology (AREA)
  • Veterinary Medicine (AREA)
  • Pathology (AREA)
  • Biophysics (AREA)
  • Dermatology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Probability & Statistics with Applications (AREA)
  • Quality & Reliability (AREA)
  • Software Systems (AREA)
  • Human Computer Interaction (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

本发明公开了一种利用皮肤图像分析皮肤水分的方法及装置。该方法中，对获得的皮肤图像进行皮肤特征分析，基于皮肤特征得到皮肤水分；其中，该皮肤特征包括光泽度和光滑度。本发明较现有技术显著降低了实施成本，还可以实现快速检测。

Description

利用皮肤图像分析皮肤水分的方法及装置 技术领域
本发明涉及一种利用皮肤图像分析皮肤水分的方法，同时也涉及一种利用皮肤图像分析皮肤水分的装置，属于皮肤检测技术领域。
背景技术
利用皮肤检测仪来测量皮肤水分的皮肤检测仪已经得到广泛应用，在皮肤美容护理以及医疗应用中受到极大的重视。
例如，专利号为ZL 201310645552.3的中国发明专利公开了一种皮肤水分检测仪。它利用多个电极贴附到人体多个部位，通过发送微小的电流信号，检测皮肤电容量分布，从而获得相应区域的皮肤水分信息。再如，申请号为201510403775.8的中国专利申请公开了一种皮肤参数测量电路与方法。它也是利用电极接触皮肤，通过水分检测模块来测量皮肤的容抗，再根据皮肤容抗与皮肤水分的关系来进行皮肤水分的检测。
由此可见，现有技术中普遍是利用传感器、电极等技术手段获得皮肤的电容或容抗等参数，再利用电容或容抗与皮肤水分之间的对应关系，得到皮肤的水分状况。
发明内容
针对现有技术的不足，本发明所要解决的首要技术问题在于提供一种利用皮肤图像分析皮肤水分的方法。
本发明所要解决的另一技术问题在于提供一种利用皮肤图像分析皮肤水分的装置。
为实现上述的发明目的，本发明采用下述的技术方案：
根据本发明实施例的第一方面，提供一种利用皮肤图像分析皮肤水分的方法，包括如下步骤：
对获得的皮肤图像进行皮肤特征分析，基于所述皮肤特征得到皮肤水分。
其中较优地，所述皮肤特征包括光泽度和光滑度。
其中较优地，基于所述皮肤特征得到皮肤水分，是在所述光泽度为 油光发亮时，判断皮肤为油性皮肤；在所述光泽度为缺少光泽时，判断皮肤为干性皮肤。
其中较优地，在所述光泽度为有光泽时，如果所述粗糙度为毛孔粗大纹理粗糙，那么判断皮肤为油性皮肤；如果所述粗糙度为较为粗糙，那么判断皮肤为干性皮肤；如果粗糙度为平滑细腻，那么判断皮肤为中性皮肤。
其中较优地，统计所述皮肤图像的亮度值L分布中，亮度大于亮度均值的像素点数量，除以皮肤图像像素点总数，得到高亮度像素点百分比；
根据预先设定的像素点百分比门限值M1和M2，其中M1大于M2，将高亮度像素点百分比大于M1的皮肤图像，确定为油光发亮；将高亮度像素点百分比小于等于M1且大于M2的皮肤图像，确定为有光泽；将高亮度像素点百分比小于等于M2的皮肤图像，确定为缺少光泽。
其中较优地，计算未经SSR算法处理的皮肤图像中的亮度值，得到第二亮度均值；根据第二亮度均值，统计所述皮肤图像中的亮度值大于所述第二亮度均值的像素点，得到第二高亮度像素点；计算所述第二高亮度像素点的个数，并计算所述第二高亮度像素点在所述皮肤图像中的百分比，得到第二高亮度像素点百分比；将所述第二高亮度像素点百分比和所述高亮度像素点百分比求和，得到百分比之和；根据所述百分比之和，与所述像素点百分比门限值M1和M2比较，如果所述百分比之和大于M1，则判断为油光发亮；如果所述百分比之和小于等于M1且大于M2，则判读为有光泽；如果所述百分比之和小于等于M2，确定为缺少光泽。
其中较优地，所述光滑度是基于皮肤粗糙度和毛孔面积均值得到的。
其中较优地，所述皮肤粗糙度的计算包括以下步骤：
步骤32：通过SSR算法将皮肤图像的对比度及亮度提高；
步骤33：通过灰度共生矩阵法算法分别计算相对方向为水平方向、对角线方向、垂直方向、反对角线方向像素间的灰度差，求出对比度，将四个方向的对比度求方差；
步骤34：利用预先设置分类用的粗糙度门限值，根据所述对比度方差值进行粗糙度分类，分为纹理粗糙、比较粗糙和细腻。
其中较优地，所述毛孔面积的计算包括以下步骤：
步骤36：将SSR算法处理后的图像进行色彩空间转换，转换为HSV色彩空间；
步骤37：在HSV色彩空间中，计算饱和度值S，将图像二值化；
步骤38：计算毛孔面积及毛孔面积均值
将所述二值化后的图像中的每一个黑白边界标记为一个圆，作为毛孔；计算所有圆面积以及圆数量，并计算其均值＝面积/圆数量，作为毛孔面积均值。
其中较优地，所述粗糙度值为平滑且所述毛孔面积均值为细腻时，判断所述皮肤图像为平滑细腻；所述粗糙度值为较为粗糙且所述毛孔面积均值为正常时，判断所述皮肤图像为比较粗糙；所述粗糙度值为纹理粗糙且所述毛孔面积均值为毛孔粗大时，判断所述皮肤图像为毛孔粗大纹理粗糙。
根据本发明实施例的第二方面，提供一种利用皮肤图像分析皮肤水分的装置，包括：
镜头模组，用于拍摄皮肤图像；
控制镜头模组的处理器，被配置为可以执行指令以实现对获得的皮肤图像进行皮肤特征分析，基于所述皮肤特征得到皮肤水分；
与处理器连接的存储器，用于存储所述处理器的可执行指令。
本发明所提供的利用皮肤图像分析皮肤水分的方法及装置，较现有技术显著降低了实施成本，并且可以实现快速检测。而且，针对没有条件使用皮肤检测仪检测的场合，利用本发明可以利用皮肤图像直接得到皮肤的水分状况，有效扩大了皮肤水分分析的使用范围。
附图说明
图1为本发明所提供的利用皮肤图像分析皮肤水分的方法流程图；
图2为图1中光泽度分析步骤的流程图；
图3为SSR算法中反射率图像与光照图像的关系示意图；
图4为SSR算法中的假设条件示意图；
图5为图2中的光泽度输出值示意图；
图6为图1中的光滑度输出值示意图；
图7为皮肤图像在二值化后显示的黑白边界示意图；
图8为图1中皮肤水分状况的定性判断步骤的流程图；
图9为本发明所提供的利用皮肤图像分析皮肤水分的装置示意图。
具体实施方式
下面结合附图和具体实施例对本发明的技术内容展开详细具体的说明。
如图1所示，本发明实施例中通过分析超过10000张不同皮肤的图像，利用皮肤图像计算分析相应的皮肤特征，再通过皮肤特征计算分析出皮肤的水分状况。所分析的皮肤特征包括皮肤的光泽度和光滑度(粗糙度)。该方法具体包括以下步骤：
步骤1：获得皮肤图像
获得皮肤图像的方式可以有多种方式，例如通过皮肤检测仪拍摄得到的皮肤图像。它可以是全脸的皮肤图像，也可以是人脸局部的皮肤图像，还可以是人体其它部分的皮肤图像。
在本步骤中，还可以进行低通滤波、侵蚀、膨胀，以去除不必要的噪声，对皮肤图像进行预处理。
步骤2：计算光泽度值
本发明是基于预处理的皮肤图像计算皮肤图像的光泽度值，得出光泽度的分类及结果。光泽度分为油光发亮、有光泽、缺少光泽。计算光泽度值的方法可以是多种，本发明实施例中采取的是图2所示的计算方法，后文将结合图2～图5进行详细说明。
步骤21：读取皮肤图像
将获取的皮肤图像进行色彩空间转换，转换为HSL色彩空间，其中L所代表的是亮度。
步骤22：通过SSR算法将图像的对比度及亮度提高。SSR算法可以将图像色彩转换成类似人眼所辨识的色彩值。
SSR(Scalable Sample Rate)算法是基于Retinex理论提出来的一种图像增强算法。基于Retinex理论，图像可以分为入射分量和反射分量，对于反射分量，使用高斯函数对该分量做卷积，然后再将两分量相加计算出最终结果。
步骤23：通过低通滤波进行图像预处理,由预处理后的图像S(x,y)取得光照图像L(x,y)，再通过以下公式算出反射率图像R(x,y)。
S(x，y)＝R(x，y)·L(x，y)
Figure PCTCN2018110130-appb-000001
r(x，y)＝log S(x，y)-log[F(x，y)*S(x，y)]
Figure PCTCN2018110130-appb-000002
步骤24：将经过SSR算法处理后的图像，分析其亮度值L的分布，得到亮度均值。
针对SSR算法处理后的图像中各个像素点的亮度值L，可以通过简单计算求得亮度均值(可以是算术平均值，也可以是几何平均值等)。
步骤25：统计大于亮度均值的像素点百分比。
统计亮度值L大于亮度均值的像素点(高亮度像素点)的个数，然后将高亮度像素点个数除以图像中的像素点总数，作为高亮度像素点百分比。
根据预先设定的像素点百分比门限值M1和M2(M1大于M2)，将高亮度像素点百分比大于M1的皮肤图像，确定为油光发亮；将高亮度像素点百分比小于等于M1且大于M2的皮肤图像，确定为有光泽；将高亮度像素点百分比小于等于M2的皮肤图像，确定为缺少光泽。其中，M1和M2是根据皮肤图像大数据统计得到的。
作为优化方案，还可以这样处理：
计算未经SSR算法处理的皮肤图像中的亮度值，得到第二亮度均值；根据第二亮度均值，统计未经SSR算法处理的皮肤图像中亮度值大于第二亮度均值的像素点，得到第二高亮度像素点；计算第二高亮度像素点的个数，并计算第二高亮度像素点在未经SSR算法处理的皮肤图像中的百分比，得到第二高亮度像素点百分比；将第二高亮度像素点百分比和高亮度像素点百分比求和，得到百分比之和；根据百分比之和，与像素点百分比门限值M1和M2比较，如果百分比之和大于M1，则判断为油光发亮；如果百分比之和小于等于M1且大于M2，则判读为有光泽；如果百分比之和小于等于M2，确定为缺少光泽。
步骤3：计算光滑度值
如图1所示，本发明是基于预处理的皮肤图像，通过计算皮肤图像中的皮肤粗糙度和毛孔面积来获得光滑度值，图6所示为光滑度的计算结果。
本发明中的光滑度值计算是基于毛孔值和纹理值获得的。其中，毛孔值和纹理值的计算均可以采用现有技术中普遍采用的各种计算方法。例如：
关于纹理值的计算，可以参考本申请人在申请号为201710337597.2的在先专利申请中公开的利用皮肤图像的颜色空间像素值计算得到偏差绝对值均值，作为皮肤图像的粗糙度特征值的方法。或者，申请号为201611197869.5的在先专利申请中公开的粗糙度提取方法，将人脸图像从RGB色彩空间转化为YCbCr色度空间，提取亮度分量、蓝色色度分量和红色色度分量图，求得肤色相似度和粗糙度。
关于毛孔值的计算，可以参考申请号为201510554895.8的在先专利申请。该专利申请公开了一种基于图像分析的皮肤毛孔识别方法，通过计算像素区域的亮度差分矩阵；融合原图像与亮度差分矩阵得到新的图像，对新图像进行聚类分析；分类迭代得到分类后的图像；统计皮肤毛孔数目，计算皮肤毛孔的平均像素面积。
本发明可以采取上述纹理值和毛孔值的计算方法，也可以采用以下计算方法分别计算粗糙度和毛孔数量。最后，将粗糙度与毛孔数量均值相结合，将皮肤光滑度划分为平滑细腻、较为粗糙、毛孔粗大纹理粗糙。
其中，计算粗糙度的步骤如下：
步骤31：获取皮肤图像
步骤32：通过SSR算法将皮肤图像的对比度及亮度提高。
步骤33：通过GLCM(灰度共生矩阵法)算法分别计算相对方向为0度(水平方向)、45度(对角线方向)、90度(垂直方向)、135度(反对角线方向)像素间的灰度差，求出对比度，将四个方向的对比度求方差。
由于沿着纹理方向上的相邻像素的灰度基本相同，垂直纹理方向上的相邻像素具有较大灰度差，因此利用主对角线元素值的大小来判断纹理走向；利用像素值的离散性，来反映纹理的粗细程度，离主对角线远的像素值高，即离散性大，表示相邻像素灰度差比例高，说明图像上垂直于该方向的纹理较细。
GLCM算法可以提取14个纹理特征，其中对比度(惯性矩)可以体现图像清晰度、纹理强弱。对比度值越大，表明纹理的沟纹越深，越容易被观察到。当前述四个方向的差异性越大时，则方差就会越大。纹理往往所呈现的是单一方向性，因此在对比度上，纹理越复杂，四个方向对比度差异越大。相反，如果皮肤平滑，则四个方向的对比度越接近，方差会越小。
步骤34：根据对比度方差值进行粗糙度分类。预先设置分类用的粗糙度门限值，根据门限值来划分粗糙度类别。
粗糙度的门限值选择是根据大量的皮肤数据统计得到的，预先存储在皮肤检测仪的存储器中。本发明实施例中设置两个粗糙度门限值，第一粗糙度门限值和第二粗糙度门限值(大于第一粗糙度门限值)。如果对比度方差值小于第一粗糙度门限值，则判断为平滑；如果对比度方差值大于第二粗糙度门限值，则判断为纹理粗糙；如果对比度方差值大于等于第一粗糙度门限值并且小于等于第二粗糙度门限值，则判断为比较粗糙。
在本发明中，基于经过SSR算法处理的皮肤图像计算毛孔面积均值的步骤包括：
步骤35：获得皮肤图像，并进行SSR算法处理
由于步骤35与计算粗糙度的步骤31和步骤32的效果一样，因此实际计算过程中可以直接在计算粗糙度时将步骤32的输出结果用于计算毛孔面积。
步骤36：将SSR算法处理后的图像进行色彩空间转换，转换为HSV色彩空间。
步骤37：在HSV色彩空间中，计算饱和度值S。
对饱和度值S，采用最大类间方差法(也称为大津法，简称OTSU)计算，计算出二值化门限值T。并利用该门限值将图像二值化，分为高饱和度区和低饱和度区两个部分。毛孔相对于皮肤皆视为灰暗部分，是低饱和度区域；皮肤是高饱和区域。越是毛孔粗大，色彩饱和度差异越明显。
最大类间方差的主要功能是找出背景与目标间的门限值。透过灰阶度的分布量，找出门限值。毛孔就是目标，皮肤则是背景。两区之间的 类间方差越大,说明构成图像的两区域的差别越大；当错分两区时会导致两区之间差别变小，因此采用最大类间方差法，得到使类间方差最大的分割门限值，就意味着错分两区的概率最小，由此可知本发明得到的毛孔计算结果较为准确。
利用最大类间方差方法找到门限值，将大于门限值(高饱和度、皮肤)转换为黑色，而小于门限值(低饱和、毛孔)转换成白色，形成黑白图像，并且在毛孔周围形成黑白边界，表示毛孔与周围皮肤的分界。
另外，本领域普通技术人员可以理解，如果对皮肤图像进行低通滤波、侵蚀、膨胀，去除不必要的噪声，可以使本发明中的毛孔计算结果更为准确。
步骤38：计算毛孔面积及毛孔面积均值
对二值化后的图像，将图像中的每一个黑白边界标记为一个圆(如图7所示)，作为毛孔。计算所有圆面积以及圆数量，并计算其均值＝面积/圆数量，作为毛孔面积均值。
皮肤检测仪的存储器内预先存储有毛孔面积门限值，将毛孔面积均值划分为毛孔粗大和毛孔正常。
步骤39：光滑度划分
如图6所示，根据前述步骤得到的粗糙度分类值和毛孔面积均值，按表1所示，将皮肤图像进行划分，得到光滑值。
Figure PCTCN2018110130-appb-000003
表1根据粗糙度和毛孔面积均值得到光滑值的表
步骤4：结合光滑度值和光泽度值计算得到皮肤水分状况
本发明所得到的皮肤水分状况是基于图像的光滑度值和光泽度值计算得到的，而不是像现有技术那样利用传感器直接检测得到。参考图8，本发明所提供的利用皮肤图像分析皮肤水分的方法中，针对皮肤水分状况的定性判断步骤如下：
步骤41：根据光泽度值，判断光泽度输出结果是否为油光。如果光泽度是油光，可以得出皮肤水分结果为油性皮肤；如果不是油光，则进入下一步；
步骤42：判断光泽度输出结果是否缺少光泽，如果光泽度是缺少光泽，可以得出皮肤水分结果为干性皮肤；如果不是缺少光泽，即皮肤为有光泽，则进入下一步；
步骤43：判断光滑度值是否为平滑细腻。如果是平滑细腻，则表示在此情况下，皮肤的光泽度是有光泽，并且光滑度是平滑细腻，那么可以得出皮肤水分结果为中性皮肤；如果光滑度不是平滑细腻，则进入下一步；
步骤44：判断光泽度值是否接近油光值。在此，接近油光值是指，在计算光泽度值时，该皮肤图像中大于高亮门限值A的像素点百分比小于但接近M1，例如该皮肤图像中大于高亮门限值A的像素点百分比为M1的85％～99％。
步骤45：如果光泽度值接近油光值，则判断为油性皮肤；否则判断为干性皮肤。换言之，利用这一步，将比较粗糙或毛孔粗大纹理粗糙的皮肤，根据光泽度是否接近油光值M1，区分为油性皮肤或干性皮肤。
本领域普通技术人员可以理解，除了利用上述方法基于光泽度和粗糙度得到水分状况，还可以利用如下表格，用查表法得到皮肤水分状况。
Figure PCTCN2018110130-appb-000004
表2基于粗糙度和光泽度的水分状况判断表
本发明利用皮肤图像分析获得皮肤水分状况，一方面避免对皮肤检测仪的依赖；另一方面，由于综合考虑了粗糙度、毛孔、光泽度多个维度，分析结果的准确性已得到实验验证。
本发明还提供一种利用皮肤图像获知皮肤水分的装置(即皮肤检测仪)。如图9所示，该装置包含具有大拍摄面积的镜头模组10，处理器20，以及与处理器20连接的存储器30。其中，镜头模组10能够实现大的拍摄面积；处理器20用于执行程序；存储器30用于存储在运行于处理器20上时能够执行各个方法步骤的程序指令，以及用于处理器运行程序时调用的门限值、水分状况判断表(例如表2)等。
由于本发明是利用皮肤图像来分析粗糙度、毛孔和水分状况，因此拍摄面积过小会导致相关皮肤特征无法被拍入一张皮肤图像中，从 而失去图像分析的可能性。为此，本发明中镜头模组的拍摄面积需要大于或等于1.5×1.5cm 2，这样能够获得足够多的皮肤特征。
上面对本发明所提供的利用皮肤图像分析皮肤水分的方法及装置进行了详细的说明。对本领域的一般技术人员而言，在不背离本发明实质精神的前提下对它所做的任何显而易见的改动，都将构成对本发明专利权的侵犯，将承担相应的法律责任。

Claims (15)

  1. 一种利用皮肤图像分析皮肤水分的方法，其特征在于：
    对获得的皮肤图像进行皮肤特征分析，基于所述皮肤特征得到皮肤水分。
  2. 如权利要求1所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    所述皮肤特征包括光泽度和光滑度。
  3. 如权利要求2所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    基于所述皮肤特征得到皮肤水分，是在所述光泽度为油光发亮时，判断皮肤为油性皮肤；在所述光泽度为缺少光泽时，判断皮肤为干性皮肤。
  4. 如权利要求3所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    在所述光泽度为有光泽时，如果所述粗糙度为毛孔粗大纹理粗糙，那么判断皮肤为油性皮肤；如果所述粗糙度为较为粗糙，那么判断皮肤为干性皮肤；如果所述粗糙度为平滑细腻，那么判断皮肤为中性皮肤。
  5. 如权利要求4所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    所述光泽度是基于SSR算法处理后的皮肤图像的亮度值分布，找出亮度均值，然后根据亮度均值判断皮肤为油光发亮、有光泽或缺少光泽。
  6. 如权利要求5所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    统计所述皮肤图像的亮度值分布中，亮度大于亮度均值的像素点数量，除以皮肤图像像素点总数，得到高亮度像素点百分比；
    根据预先设定的像素点百分比门限值M1和M2，其中M1大于M2，将高亮度像素点百分比大于M1的皮肤图像，确定为油光发亮；将高亮度像素点百分比小于等于M1且大于M2的皮肤图像，确定为有光泽；将高亮度像素点百分比小于等于M2的皮肤图像，确定为缺少光泽。
  7. 如权利要求6所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    计算未经SSR算法处理的皮肤图像的亮度值，得到第二亮度均值；
    根据第二亮度均值，统计所述皮肤图像中的亮度值大于所述第二亮度均值的像素点，得到第二高亮度像素点；
    计算所述第二高亮度像素点的个数，并计算所述第二高亮度像素点在所述皮肤图像中的百分比，得到第二高亮度像素点百分比；
    将所述第二高亮度像素点百分比和所述高亮度像素点百分比求和，得到百分比之和；
    根据所述百分比之和，与所述像素点百分比门限值M1和M2比较，如果所述百分比之和大于M1，则判断为油光发亮；如果所述百分比之和小于等于M1且大于M2，则判读为有光泽；如果所述百分比之和小于等于M2，确定为缺少光泽。
  8. 如权利要求2所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    所述光滑度是基于皮肤粗糙度和毛孔面积均值得到的。
  9. 如权利要求8所述的利用皮肤图像分析皮肤水分的方法，其特征在于所述皮肤粗糙度的计算包括以下步骤：
    步骤32：通过SSR算法将皮肤图像的对比度及亮度提高；
    步骤33：通过灰度共生矩阵法算法分别计算相对方向为水平方向、对角线方向、垂直方向、反对角线方向像素间的灰度差，求出对比度，将四个方向的对比度求方差；
    步骤34：利用预先设置分类用的粗糙度门限值，根据所述对比度方差值进行粗糙度分类，分为纹理粗糙、比较粗糙和细腻。
  10. 如权利要求8所述的利用皮肤图像分析皮肤水分的方法，其特征在于所述毛孔面积的计算包括以下步骤：
    步骤36：将SSR算法处理后的图像进行色彩空间转换，转换为HSV色彩空间；
    步骤37：在HSV色彩空间中，计算饱和度值S，将图像二值化；
    步骤38：计算毛孔面积及毛孔面积均值：将所述二值化后的图像中的每一个黑白边界标记为一个圆，作为毛孔；计算所有圆面积以及圆数 量，并计算其均值＝面积/圆数量，作为毛孔面积均值。
  11. 如权利要求9或10所述的利用皮肤图像分析皮肤水分的方法，其特征在于：
    所述粗糙度值为平滑且所述毛孔面积均值为细腻时，判断所述皮肤图像为平滑细腻；所述粗糙度值为较为粗糙且所述毛孔面积均值为正常时，判断所述皮肤图像为比较粗糙；所述粗糙度值为纹理粗糙且所述毛孔面积均值为毛孔粗大时，判断所述皮肤图像为毛孔粗大纹理粗糙。
  12. 一种利用皮肤图像分析皮肤水分的装置，其特征在于包括：
    镜头模组，用于拍摄大面积的皮肤图像；
    控制镜头模组的处理器，被配置为可以执行指令以实现对获得的皮肤图像进行皮肤特征分析，基于所述皮肤特征得到皮肤水分；
    与处理器连接的存储器，用于存储处理器可执行指令。
  13. 如权利要求12所述的利用皮肤图像分析皮肤水分的装置，其特征在于：
    所述皮肤特征包括光泽度和光滑度。
  14. 如权利要求13所述的利用皮肤图像分析皮肤水分的装置，其特征在于：
    基于所述皮肤特征得到皮肤水分，是在所述光泽度为油光发亮时，判断皮肤为油性皮肤；在所述光泽度为缺少光泽时，判断皮肤为干性皮肤。
  15. 如权利要求14所述的利用皮肤图像分析皮肤水分的装置，其特征在于：
    在所述光泽度为有光泽时，如果所述粗糙度为毛孔粗大纹理粗糙，那么判断皮肤为油性皮肤；如果所述粗糙度为较为粗糙，那么判断皮肤为干性皮肤；如果所述粗糙度为平滑细腻，那么判断皮肤为中性皮肤。
PCT/CN2018/110130 2017-10-12 2018-10-12 利用皮肤图像分析皮肤水分的方法及装置 WO2019072249A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/846,326 US11490854B2 (en) 2017-10-12 2020-04-11 Method and device for analyzing water content of skin by means of skin image

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710948026.2 2017-10-12
CN201710948026.2A CN109671046B (zh) 2017-10-12 2017-10-12 利用皮肤图像分析皮肤水分的方法及装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/846,326 Continuation US11490854B2 (en) 2017-10-12 2020-04-11 Method and device for analyzing water content of skin by means of skin image

Publications (1)

Publication Number Publication Date
WO2019072249A1 true WO2019072249A1 (zh) 2019-04-18

Family

ID=66100415

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/110130 WO2019072249A1 (zh) 2017-10-12 2018-10-12 利用皮肤图像分析皮肤水分的方法及装置

Country Status (3)

Country Link
US (1) US11490854B2 (zh)
CN (1) CN109671046B (zh)
WO (1) WO2019072249A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115063566A (zh) * 2022-08-15 2022-09-16 天津艺点意创科技有限公司 一种基于ar的文创产品展示方法及展示设备
EP4030343A4 (en) * 2019-10-22 2022-10-26 Huawei Technologies Co., Ltd. FACIAL SKIN DETECTION METHOD AND APPARATUS
CN116777919A (zh) * 2023-08-25 2023-09-19 湖南云天工程检测有限公司 一种混凝土试件智能养护方法及系统

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20210106990A (ko) * 2018-11-30 2021-08-31 피씨엠에스 홀딩스, 인크. 피부 반사율 데이터베이스에 기초하여 장면 발광체를 추정하는 방법 및 장치
CN110057325B (zh) * 2019-04-26 2020-06-23 湖南大学 一种基于成像仿真的表面粗糙度检测方法及计算设备
CN110728247A (zh) * 2019-10-18 2020-01-24 南京农业大学 一种猪胴体表皮有水或无水状态识别方法
CN113487562B (zh) * 2021-07-02 2024-04-09 北京美医医学技术研究院有限公司 一种基于手持设备的皮肤光泽度检测系统
CN116912235B (zh) * 2023-09-08 2023-11-24 金乡县金禧园食品有限公司 基于计算机视觉的脱水蒜状态智能检测方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN104732214A (zh) * 2015-03-24 2015-06-24 吴亮 一种基于人脸图像识别的定量化肤质检测方法
CN104732200A (zh) * 2015-01-28 2015-06-24 广州远信网络科技发展有限公司 一种皮肤类型和皮肤问题的识别方法
CN106469243A (zh) * 2015-08-14 2017-03-01 集粹坊科贸（北京）有限责任公司 一种基于数字图像处理技术的人体面部皮肤水分、弹性指标计算方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3054630B1 (en) * 2005-04-15 2019-01-23 InterDigital CE Patent Holdings Remote management method of a device and corresponding device
JP6030396B2 (ja) * 2012-09-28 2016-11-24 日立マクセル株式会社 映像処理装置
WO2014196532A1 (ja) * 2013-06-07 2014-12-11 富士フイルム株式会社 透明感評価装置、透明感評価方法および透明感評価プログラム
CN105069818A (zh) * 2015-09-02 2015-11-18 泰山学院 一种基于图像分析的皮肤毛孔识别方法
CN107169960B (zh) * 2017-05-15 2020-03-20 精诚工坊电子集成技术（北京）有限公司 一种基于图像颜色空间的皮肤表面毛孔大小检测方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN104732200A (zh) * 2015-01-28 2015-06-24 广州远信网络科技发展有限公司 一种皮肤类型和皮肤问题的识别方法
CN104732214A (zh) * 2015-03-24 2015-06-24 吴亮 一种基于人脸图像识别的定量化肤质检测方法
CN106469243A (zh) * 2015-08-14 2017-03-01 集粹坊科贸（北京）有限责任公司 一种基于数字图像处理技术的人体面部皮肤水分、弹性指标计算方法

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4030343A4 (en) * 2019-10-22 2022-10-26 Huawei Technologies Co., Ltd. FACIAL SKIN DETECTION METHOD AND APPARATUS
CN115063566A (zh) * 2022-08-15 2022-09-16 天津艺点意创科技有限公司 一种基于ar的文创产品展示方法及展示设备
CN115063566B (zh) * 2022-08-15 2022-11-22 天津艺点意创科技有限公司 一种基于ar的文创产品展示方法及展示设备
CN116777919A (zh) * 2023-08-25 2023-09-19 湖南云天工程检测有限公司 一种混凝土试件智能养护方法及系统
CN116777919B (zh) * 2023-08-25 2023-11-07 湖南云天工程检测有限公司 一种混凝土试件智能养护方法及系统

Also Published As

Publication number Publication date
CN109671046B (zh) 2021-08-17
US11490854B2 (en) 2022-11-08
CN109671046A (zh) 2019-04-23
US20200237286A1 (en) 2020-07-30

Similar Documents

Publication Publication Date Title
WO2019072249A1 (zh) 利用皮肤图像分析皮肤水分的方法及装置
Al-Musawi et al. Three-phase induction motor fault detection based on thermal image segmentation
WO2019062092A1 (zh) 基于超像素-多色彩空间的人体轮廓提取方法
CN110097034A (zh) 一种智能人脸健康程度的识别与评估方法
CN105787929B (zh) 基于斑点检测的皮肤疹点提取方法
Ren et al. Fusion of intensity and inter-component chromatic difference for effective and robust colour edge detection
JP5706233B2 (ja) 鋼材成分識別装置及びそのプログラム
CN116758077B (zh) 一种冲浪板表面平整度在线检测方法及系统
CN113935666B (zh) 基于图像处理的建筑装饰墙面砖异常评估方法
CN105787912B (zh) 一种基于分类的阶跃型边缘亚像素定位方法
KR101769741B1 (ko) 동공 탐지를 통한 홍채 인식 방법 및 홍채 인식 장치
TWI586954B (zh) 檢測細胞受人類乳突病毒（ｈｐｖ）感染的裝置及其檢測方法
Yusuf et al. Human face detection using skin color segmentation and watershed algorithm
JP2006061170A (ja) 皮膚の鑑別法
CN113706515B (zh) 舌像异常确定方法、装置、计算机设备和存储介质
Li et al. A perceptual color edge detection algorithm
CN110532993B (zh) 一种人脸防伪方法、装置、电子设备及介质
JP2009258770A (ja) 画像処理方法、画像処理装置、画像処理プログラム、撮像装置
KR100927642B1 (ko) 영상 분할 히스토그램 분석을 이용한 얼굴 윤곽선 검출 방법
CN111523605A (zh) 一种图像识别方法、装置、电子设备及介质
TWI602155B (zh) 利用影像內容不連續性增強物件偵測之方法
CN108805855A (zh) 一种混凝土裂纹快速识别方法
KR100882139B1 (ko) 얼굴 검출 방법
KR20100094045A (ko) 유두와 가슴선을 이용한 여성의 가슴 탐지 방법
CN117115174B (zh) 一种钳子外形自动检测方法及系统

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: 18867264

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: 18867264

Country of ref document: EP

Kind code of ref document: A1