Hair image processing method, system, computer device, medium and program product

Abstract

This application relates to a hair image processing method, apparatus, computer device, storage medium, and computer program product. The method comprises: acquiring collected hair images corresponding to different attribute values ​​for the same attribute; calculating the differences between the hair images corresponding to the different attribute values ​​to obtain a difference image; and performing hair recognition on the difference image to obtain a hair recognition result. Using this method, the background region absorbs light of different attributes differently, while the hair absorbs light of different attribute values ​​substantially uniformly. Therefore, segmenting and recognizing hair using this difference image can improve the stability of the hair recognition result.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a hair image processing method, system, computer device, storage medium and computer program product. Background Art

[0002] Hair transplantation involves surgically redistributing some of the remaining hair from the back of the scalp to areas of hair loss on the scalp or other areas of the body where hair has been lost. The transplanted hair retains all of its original growth characteristics and continues to grow in the new transplanted area, remaining there for life. The results are generally long-lasting. Hair transplantation is currently one of the most effective technologies for improving permanent hair loss.

[0003] Before hair transplantation, it is often necessary to study the current state of hair growth. Traditionally, robots have been used to capture this information. Specifically, when using robots for automated hair extraction, the robot's vision system must perform real-time recognition and segmentation of hairs. However, current methods process a single image, which is easily affected by the environment and can lead to inaccurate recognition and segmentation. Summary of the Invention

[0004] Based on this, it is necessary to provide a hair image processing method, system, computer device, computer-readable storage medium and computer program product that can improve the stability of recognition and segmentation effects in order to address the above technical problems.

[0005] In a first aspect, the present application provides a hair image processing method, the method comprising:

[0006] Obtaining hair images corresponding to different attribute values ​​under the same attribute;

[0007] Calculate the difference of hair images corresponding to different attribute values ​​to obtain a difference image;

[0008] Performing hair recognition on the difference image to obtain a hair recognition result.

[0009] In one embodiment, the acquiring of the collected hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following:

[0010] Using a camera to photograph the same hair area under different light sources to obtain hair images corresponding to different attribute values ​​under the same attribute; or

[0011] The same hair area is photographed by different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0012] In one embodiment, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following:

[0013] Changing the property value of the light source, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different property values ​​under the same property; or

[0014] Turn on different light sources in sequence, and use a camera to shoot the same hair area under different light sources to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0015] In one embodiment, changing the property value of the light source includes:

[0016] The property value of the light source output is changed, or the property value of the light source is changed through a filter.

[0017] In one embodiment, photographing the same hair area with different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following:

[0018] The same hair area is photographed by different cameras to obtain a two-dimensional hair image or a depth hair image corresponding to different attribute values ​​under the same attribute; or the same hair area is photographed by different cameras to obtain a two-dimensional hair image and a depth hair image respectively, and the attributes of the two-dimensional hair image and the depth hair image are the same, but the attribute values ​​are different.

[0019] In one embodiment, photographing the same hair area with different cameras to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute includes:

[0020] Different cameras are used to capture images formed by the reflector after reflection in different directions to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute. The reflector is used to reflect light passing through the same hair area.

[0021] In one embodiment, calculating the difference in hair images corresponding to different attribute values ​​under the same attribute to obtain the difference image includes:

[0022] Obtaining pre-calibrated projection relationships of different cameras;

[0023] Projecting the depth hair image onto the image plane of the two-dimensional hair image according to the projection relationship to obtain an image to be processed;

[0024] The difference between the image to be processed and the two-dimensional hair image is calculated to obtain a difference image.

[0025] In one embodiment, calculating the difference in hair images corresponding to different attribute values ​​under the same attribute to obtain a difference image includes:

[0026] Obtaining a projection relationship between two cameras in a pre-calibrated binocular camera, wherein the binocular camera includes a monocular camera and a natural light camera;

[0027] Projecting the hair image captured by one of the cameras onto the image plane of the hair image captured by the other camera according to the projection relationship to obtain an image to be processed; or

[0028] The difference between the image to be processed and the hair image captured by another camera is calculated to obtain a difference image.

[0029] In one embodiment, performing hair recognition on the difference image to obtain a hair recognition result includes:

[0030] performing threshold segmentation on the difference image to obtain a hair region;

[0031] Contour extraction is performed on the hair area to obtain a hair recognition result.

[0032] In one embodiment, after extracting the contour of the hair area to obtain the hair recognition result, the method includes:

[0033] Hair attributes are obtained based on the hair recognition results, where the hair attributes include at least one of hair quantity, hair thickness, and hair length.

[0034] In a second aspect, the present application further provides a hair recognition system, comprising:

[0035] A light source module, configured to provide a light source for irradiating the hair;

[0036] A camera module is used to collect hair images corresponding to different attribute values ​​under the same attribute;

[0037] The processing module is used to process the collected hair images corresponding to different attribute values ​​under the same attribute according to the above-mentioned hair image processing method to obtain hair recognition results.

[0038] In one embodiment, the light source module includes a light source with adjustable properties, or the light source module includes an adjustable filter.

[0039] In one embodiment, the camera module includes different types of two-dimensional cameras, or includes different types of cameras for collecting depth information, or includes a camera for collecting depth information and a two-dimensional camera. In one embodiment, the camera module also includes a reflector configured to reflect light from the light source module toward different cameras.

[0040] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in any one of the above embodiments when executing the computer program.

[0041] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method in any one of the above-mentioned embodiments when the computer program is executed by a processor.

[0042] In a fifth aspect, the present application also provides a computer program product, comprising a computer program, which implements the steps of the method in any one of the above embodiments when executed by a processor.

[0043] The above-mentioned hair image processing method, apparatus, computer device, storage medium and computer program product collect hair images corresponding to different attribute values ​​under the same attribute to calculate a difference image. Because the background area absorbs light of different attributes differently, while hair absorbs light of different attribute values ​​in a basically consistent manner, segmentation and recognition using this difference image can improve the accuracy of hair recognition results. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a schematic diagram of a hair image processing system according to an embodiment;

[0045] Figure 2 FIG1 is a hardware framework diagram of a hair image processing system in one embodiment;

[0046] Figure 3 is a schematic diagram of a camera module and a light source module in one embodiment;

[0047] Figure 4 is a schematic diagram of a camera module and a light source module in yet another embodiment;

[0048] Figure 5 is a schematic diagram of a camera module and a light source module in yet another embodiment;

[0049] Figure 6 1 is a flow chart of a hair image processing method according to an embodiment;

[0050] Figure 7is a schematic diagram of a display interface in an embodiment;

[0051] Figure 8 is a flowchart of a hair image processing method in another embodiment;

[0052] Figure 9 is a flowchart of a hair image processing method in yet another embodiment;

[0053] Figure 10 is a flowchart of a hair image processing method in yet another embodiment;

[0054] Figure 11 is a schematic diagram of a natural light image in one embodiment;

[0055] Figure 12 is a schematic diagram of a depth image in one embodiment;

[0056] Figure 13 is a schematic diagram of the result after threshold segmentation is performed on the difference image in one embodiment;

[0057] Figure 14 is a schematic diagram after contour extraction in one embodiment;

[0058] Figure 15 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0060] The hair image processing method provided in the embodiment of the present application can be applied to Figure 1 The hair image processing system shown in FIG. 1 includes a light source module 200, a camera module 300, and a processing module 100. The light source module 200 and the camera module 300 are in communication with the processing module 100, respectively. The camera module 300 and the light source module 200 can be connected to the processing module 100 via a data cable.

[0061] Specifically, combined Figure 1The camera module 300 and light source module 200 are fixedly connected, and the entire device is mounted on a robotic arm 400. In other embodiments, the device can also be mounted on any fixed bracket. During use, the patient sits in a chair and exposes their head. The doctor moves the camera module 300 near the patient's head and begins to capture and process hair images. In other embodiments, the camera module 300 and light source module 200 can be moved manually or through control commands issued by the processing module 100, without specific limitation.

[0062] Combine Figure 2 As shown, Figure 2 The hardware framework diagram of a hair image processing system in one embodiment includes a light source module 200 for providing light to illuminate hair, a camera module 300 for capturing hair images corresponding to different attribute values ​​under the same attribute, and a processing module 100 for processing the captured hair images to obtain hair recognition results. In other embodiments, the system further includes a display 500, which is connected to the processing module 100 and is used to display the hair recognition results. In other embodiments, the processing module 100 and the display 500 can be integrated into a terminal or server, without specific limitation herein.

[0063] Specifically, the doctor moves the camera near the patient's head, so that the camera module 300 captures an image of the hair area and sends it to the processing module 100. The processing module 100 can control the light source module 200 to change, control the camera module 300 to capture hair images again, and process the captured hair images to obtain hair recognition results. The light source module 200 can change the property values ​​of the output light source according to the instructions of the processing module 100, such as changing the color or intensity. The display 500 can receive the hair recognition results sent by the processing module 100 for display. In addition, the display 500 can also display the hair images captured by the camera module 300. For example, at the beginning, the hair image displayed on the display 500 can be used to determine whether the camera has moved to the target position.

[0064] In order to enable those skilled in the art to fully understand the present solution, and to obtain hair images corresponding to different attribute values ​​under the same attribute, the present application processes the images in two ways: one is through different light sources, and the other is through different types of cameras.

[0065] Specifically, combined Figures 3 to 5 As shown, Figures 3 to 5 are schematic diagrams of a camera module 300 and a light source module 200 in an embodiment, and the following describes Figure 3 and Figure 5 To explain:

[0066] like Figure 3 As shown, in one embodiment, the light source module 200 includes a light source with adjustable properties, or the light source module 200 includes an adjustable filter. The property values ​​of the light source module 200, such as color and / or intensity, can be changed. Specifically, the light source module 200 can be a controllable and adjustable illumination light color and / or intensity, or the light source module 200 includes a controllable filter, so as to control the property value of the light source output by controlling the filter. The camera module 300 can be a camera module 300 that can capture images in real time at a high frame rate. The camera can use a two-dimensional camera or a camera that collects depth information or other imaging devices, wherein the two-dimensional camera can be a monocular camera, and the camera that collects depth information can be a binocular camera or a depth camera, which is not specifically limited here. Combined with Figure 3 The light source module 200 can be a light strip, an annular light source, a backlight source, a bar light source, a coaxial light source, a multi-angle shadowless light source, or a point light source. The light source module 200 provides illumination with different attribute values ​​according to the instructions of the processing module 100. When using a filter, the filter can be controlled by the processing module 100.

[0067] In this embodiment, the processor controls the property values ​​of the light source module 200 and the camera module 300 to obtain hair images with different property values ​​under the same property, wherein the property refers to light color and / or intensity.

[0068] like Figure 4 As shown, in one embodiment, the camera module 300 includes different types of two-dimensional cameras, or includes different types of cameras that collect depth information, or includes both a two-dimensional camera and a camera that collects depth information. Specifically, the camera module 300 may include a camera combination consisting of at least two cameras, so that the camera combination can capture the same hair area to obtain hair images with different attribute values ​​under the same attribute. In this embodiment, the cameras in the camera combination can all be two-dimensional cameras, or all be cameras that collect depth information, or one can be a two-dimensional camera and the other a camera that collects depth information. Specifically, the camera module 300 includes a camera combination and a reflector 600. Specifically, the reflector 600 can reflect light passing through the same hair area in different directions, so that the cameras in the camera combination can capture images of the same hair area.

[0069] Regarding the camera combination, when the light source module 200 is an infrared light source, the camera combination can be a monocular infrared camera capable of real-time shooting at a high frame rate and a monocular natural light camera capable of real-time shooting at a high frame rate; or a binocular infrared camera and a monocular natural light camera capable of real-time shooting at a high frame rate; or a monocular infrared camera capable of real-time shooting at a high frame rate and a binocular natural light camera; or a monocular infrared camera capable of real-time shooting at a high frame rate and a binocular natural light camera; or a monocular infrared camera capable of real-time shooting at a high frame rate and a binocular natural light camera; or a binocular black and white camera and a binocular natural light camera; or a binocular black and white camera and a binocular natural light camera. In other embodiments, the camera combination can also be any two cameras that can capture different effects on the same hair area.

[0070] In this embodiment, the processing module 100 controls different cameras to collect hair images with the same attribute but different attribute values ​​under the same light source.

[0071] like Figure 5 As shown, in one embodiment, the camera module 300 includes at least one camera for collecting depth information and at least one two-dimensional camera. The camera module 300 may include a camera combination consisting of at least two cameras, so that the camera combination can capture the same hair area to obtain hair images with different attribute values ​​under the same attribute. The camera combination may include a camera for collecting depth information and a two-dimensional camera. In other embodiments, the camera combination may include a monocular camera and a natural light camera, and the two cameras are pre-calibrated to obtain a projection relationship of the image captured by one camera onto the image plane captured by the other camera. When the camera captures the hair area, which can be approximately regarded as a plane area, the image captured by one camera of the binocular camera is projected onto the image plane captured by the other camera, thereby obtaining hair images with different attribute values ​​under the same attribute in the same plane.

[0072] like Figure 5 , wherein the two-dimensional camera in the camera combination can be a natural light camera, and the camera for collecting depth information can be obtained by combining a natural light camera and a monocular structured light camera. In other embodiments, it can also be composed of any imaging system and a monocular structured light camera, which is not specifically limited here.

[0073] In this embodiment, the processing module 100 controls different cameras to capture a two-dimensional hair image and a depth image under the same light source, and then converts the depth image into a two-dimensional hair image plane through a pre-calibrated projection relationship, so that the projected image to be processed and the two-dimensional hair image are hair images with different attribute values ​​under the same attributes.

[0074] In one embodiment, Figure 6 As shown, a hair image processing method is provided, which is applied to Figure 1 The processing module 100 in FIG. 1 is taken as an example to illustrate the process, which includes the following steps:

[0075] S602: Acquire collected hair images corresponding to different attribute values ​​under the same attribute.

[0076] Specifically, attributes may include color and intensity, where the attribute value refers to different colors or different intensities. Different colors or different intensities can be achieved by controlling changes in the light source module or by using different cameras. Changes in the light source module include changes to the light source itself or changes to the filter.

[0077] In actual applications, the processing module 100 can be controlled to capture hair images of different colors or different intensities. Specifically, the processing module 100 can control the light source module 200 to emit light of different attribute values ​​under the same attribute at different times. In other embodiments, the processing module 100 can control different cameras to capture hair areas under the same light source to obtain hair images with different attribute values ​​under the same attribute.

[0078] S604: Calculate the difference between the hair images corresponding to different attribute values ​​to obtain a difference image.

[0079] Specifically, the difference image is obtained based on the difference in hair images corresponding to different attribute values. The difference can be calculated by calculating the difference in pixel values ​​at corresponding positions of the hair image, where the difference can be the difference or differential of the pixel values ​​at corresponding positions, etc., which is not specifically limited here.

[0080] It should be noted that the corresponding position refers to the same position in the hair region, that is, the position of the same hair region in different hair images. When the attribute value of the light source module 200 is changed, since the same camera is used for shooting, the camera position remains unchanged, so the pixel difference can be directly calculated for the pixels with the same coordinates in different hair images. When different cameras are used for shooting, if the reflector 600 is introduced, since the reflector 600 reflects the light passing through the same hair region, the hair regions captured by different cameras are exactly the same, so the pixel difference can be directly calculated for the pixels with the same coordinates in different hair images. If the reflector 600 is not introduced, since the different cameras have perspective deviations, they can be converted to the same plane based on pre-calibration, and then, based on the pre-calibrated perspective deviation, the pixel difference can be calculated only for the pixels at corresponding positions in the hair images of the cameras' common viewing area.

[0081] In practical applications, the processing module 100 can represent the hair image using a pixel matrix. By calculating the difference between the matrices, the difference between the hair images can be obtained. The obtained difference matrix corresponds to the difference image. Specifically, the processing module 100 can directly calculate the difference between the matrices to obtain a difference matrix, and use this difference matrix to represent the difference image.

[0082] S606: Perform hair recognition on the difference image to obtain a hair recognition result.

[0083] The hair recognition result refers to the area of ​​hair, etc. In one embodiment, after performing hair recognition on the difference image to obtain the hair recognition result, it includes: obtaining hair attributes based on the hair recognition result, and the hair attributes include at least one of the number of hairs, hair thickness and hair length.

[0084] Specifically, combined Figure 7 As shown, the processing module 100 can send the hair recognition results and hair attributes to the display module 500 for display. The number of hairs can be determined by extracting the contours of the hair region in the difference image, and the number of extracted contours is the number of hairs. The thickness of the hair can be calculated from the short side of the extracted contour, and the length of the hair can be calculated from the long side of the extracted contour.

[0085] In one embodiment, the number of hairs can be determined by simply identifying the difference image. For hair thickness and length, the processing module 100 can reproject the identified hair contours into the depth image to determine the hair region in the depth image, and then calculate the hair thickness and / or hair length based on the data in the depth image. In other embodiments, the processing module 100 can determine the conversion relationship between pixel units and length units, such as millimeters, based on the mapping relationship between the depth image and the two-dimensional image. In this way, the pixel length of the short side of the contour in the difference image can be read, and the hair thickness in millimeters can be obtained based on this conversion relationship. Similarly, the calculation of hair length is similar and will not be repeated here. In one embodiment, hair thickness and / or hair length can be identified by statistical values ​​of hair thickness and / or hair length in the hair region, such as the average value of hair thickness and / or hair length, and the average value is output for the doctor's reference.

[0086] In one embodiment, the hair is black. Specifically, non-black hair can be dyed black before surgery. Black hair absorbs all colors of light, and since the scalp is not black, it absorbs different colors of light differently. This difference can be used to segment hair regions. Specifically, black hair's absorption of light sources is reflected by smaller grayscale value variations in images of different light sources, while the grayscale value variations of the scalp region under different light sources are larger. Therefore, hair regions can be segmented based on this hair feature. Furthermore, due to the diffuse reflection characteristics of hair, it does not strongly reflect light when illuminated, and thus, there are no areas of intense exposure, allowing the aforementioned grayscale value variations to be accurately identified. This black characteristic of hair reduces the impact of ambient light on hair segmentation and counting.

[0087] The above-mentioned hair image processing method collects hair images corresponding to different attribute values ​​under the same attribute and calculates a difference image. Since the scalp area absorbs light of different attributes differently, while the hair absorbs light of different attribute values ​​in a basically consistent manner, segmentation and recognition using this difference image can improve the stability of the hair recognition results. In addition, the above-mentioned hair image processing method improves the speed of automatic hair extraction and reduces the operation time by improving the calculation speed of the hair segmentation method. By improving the anti-interference ability of hair segmentation, the accuracy of hair positioning is improved, and the accuracy of hair extraction is improved. The computational complexity of hair segmentation is reduced, and the number of times greater than N is increased. 2The computational complexity is reduced to 2N, reducing the heat generated by the machine during calculations. This hair image processing method is applicable to different ethnic groups, avoiding significant differences in segmentation results for different ethnic groups. Because the calculation time does not increase significantly with the increase in resolution, it opens up the possibility of using high-resolution cameras for hair segmentation. By improving the efficiency and anti-interference performance of hair segmentation, the number of hairs in the image can be quickly counted. At the same time, when using a camera that collects depth information, it can also provide the thickness of large areas of hair in the image to assist doctors in diagnosis.

[0088] In one embodiment, obtaining collected hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following: photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute; or photographing the same hair area with different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0089] Specifically, the light source module can be changed by changing the property value of the light source or turning on a different light source. Changing the property value of the light source can include changing the property value output by the light source or changing the property value through a filter.

[0090] When changing the light source, the light source can be changed at least once, for example, by changing the attribute value of the light emitted by the light source at least once via a command, or by sequentially activating at least two different light sources. The difference image can be obtained by calculating the difference between the two hair images, and then calculating the corresponding statistics, such as the average, to obtain the difference image. In one embodiment, multiple hair images taken at different times can be captured under a light source with the same attribute value. Statistics, such as the mean, of the multiple hair images taken at different times can be calculated to obtain the hair image corresponding to the same attribute value. The light source can then be changed to obtain at least one hair image with the same attribute but different attribute value. The difference image can be obtained by averaging the difference between the two hair images.

[0091] In the case of using different cameras, the processing module 100 can calculate a statistical quantity, such as a mean value, of the differences between the hair images captured by each camera to obtain a difference image. In one embodiment, the same camera can capture multiple hair images at different times, and then obtain a hair image of one camera by calculating a statistical quantity, such as the mean value of the hair images captured by the same camera at different times. The difference image is then calculated using the hair images from at least two cameras. Alternatively, the difference between hair images captured by different cameras at the same time is calculated, and then a statistical quantity, such as a mean value, is calculated for the differences. Finally, a statistical quantity, such as a mean value, is calculated for the differences between at least two times to obtain a difference value, thereby obtaining a difference image.

[0092] In the above embodiments, the attribute values ​​of the hair image are changed in different ways, thereby improving universality.

[0093] For those skilled in the art to fully understand, the above two methods are introduced separately, mainly including three different implementation methods:

[0094] The first approach combines controllable and adjustable multi-color and / or intensity ambient light bands or through adjustable filters with a high frame rate monocular or binocular camera that can capture images in real time.

[0095] The second method is to use a reflector 600 structure that can reflect the same image to two cameras, and use a combination of an infrared camera and a natural light camera or a combination of a natural light and a black and white light camera to capture images of the same area to obtain a combination of images of different colors.

[0096] The third method is to use two calibrated cameras, which can be a combination of an infrared camera and a natural light camera, or a natural light and black and white light camera to capture images in different colors.

[0097] In other embodiments, the color attribute may be changed to an intensity attribute, which is not specifically limited herein.

[0098] In one embodiment, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following: changing the attribute value of the light source, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute; or turning on different light sources in sequence, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0099] In one embodiment, changing the property value of the light source includes: changing the property value output by the light source, or changing the property value of the light source through a filter.

[0100] In one embodiment, in combination with the above Figure 3 、 Figure 8 As shown, Figure 8 FIG. 4 is a flowchart of a hair image processing method in another embodiment.

[0101] In this embodiment, the property value of the light source can be changed by changing the output of the light source, or by turning on a light source with a different property value, and the output of the light source can be changed by changing the property value of the light source output, or by changing the property value of the light source through a filter. In this embodiment, this is mainly achieved through a controllable and adjustable light strip. For details, please refer to Figure 3 ,in Figure 3In the embodiment, a binocular camera is used, and in other embodiments, a monocular camera can be used. In addition, in other embodiments, the output of the light source can be changed by other means, which are not specifically limited here.

[0102] At the beginning of the operation, the camera is turned on, and the doctor moves the camera to the target position. The camera captures the hair area to obtain Image 1, and the captured image is returned to the processing module 100. After receiving Image 1, the processing module 100 changes the color or intensity of the light source and sends a command to the camera module 300 to take a second shot. The camera captures the hair area to obtain Image 2. Since the position of the camera does not move during the entire process, Image 1 and Image 2 are captured for the same hair area.

[0103] Processing module 100 calculates the difference between image 1 and image 2 to obtain a difference image. Specifically, it subtracts the pixel value matrices of image 1 and image 2 to obtain a difference matrix. Based on the difference matrix, it generates a grayscale image. Threshold segmentation is then performed on the grayscale image to determine the hair region and scalp region. The hair region is then contoured to obtain the outline of each hair, thereby obtaining a representation of all hairs in the image. The number of contours is counted to determine the number of hairs in the image.

[0104] In practical applications, the camera first captures the hair region to obtain a natural light image; then, the color and / or intensity of the light source are changed, and the hair region is captured again to obtain a new image. The captured images show that the color and / or intensity of the hair in the two images are similar, while the color and / or intensity of the scalp region vary significantly. When the two images are subtracted, the region with the smaller value represents the hair region. Therefore, threshold segmentation can be used to determine the hair region. This threshold can be determined empirically and is not specifically defined here. Because hair and scalp absorb light differently, the hair and surrounding areas, such as the scalp, need to be set to different colors. In one embodiment, the hair is black. Therefore, the processing module 100 can set a threshold based on the color of the hair and surrounding areas to achieve accurate segmentation of the hair region.

[0105] In the above embodiment, the light source is changed to obtain hair images with different attribute values ​​for the same attribute, and then segmentation is performed, thereby improving efficiency.

[0106] In one embodiment, photographing the same hair area with different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following: photographing the same hair area with different cameras to obtain two-dimensional hair images corresponding to different attribute values ​​under the same attribute; or photographing the same hair area with different cameras to obtain depth hair images corresponding to different attribute values ​​under the same attribute; or photographing the same hair area with different cameras to obtain a two-dimensional hair image and a depth hair image respectively, and the two-dimensional hair image and the depth hair image have the same attribute but different attribute values.

[0107] Specifically, this embodiment primarily changes attribute values ​​by changing cameras. This can be categorized into two types: one where different cameras share the same category, such as both being 2D cameras or both being cameras that collect depth information. The other type is a camera combination that includes at least one 2D camera and at least one camera that collects depth information.

[0108] Specifically, when hair images are collected through different cameras of the same category, the same hair area is photographed through different cameras to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute, including: using different cameras, collecting images formed after the reflector 600 reflects in different directions to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute, and the reflector 600 is used to reflect light passing through the same hair area.

[0109] In one embodiment, in combination with the above Figure 4 、 Figure 9 As shown, Figure 9 4 is a flowchart of a hair image processing method in yet another embodiment.

[0110] In this embodiment, it is assumed that the different cameras are an infrared camera and a natural light camera. In other embodiments, other different cameras may be used, as described above. The infrared camera and the natural light camera are activated by the processing module 100. The doctor moves the camera to a designated area, and the processing module 100 controls the turning on of the infrared light of the light source module. In other embodiments, the light source module may also use other light sources, such as a conventional light source, etc., which will not be described in detail here.

[0111] The infrared camera and the natural light camera capture images of the same hair area from the same viewing angle via a reflector 600, generating images 1 and 2. These images are then returned to the processing module 100, which subtracts the pixel matrices of images 1 and 2 to obtain a difference image matrix. This difference image matrix is ​​then used to generate a difference image. This difference image is then thresholded to identify the hair area and scalp area. The hair area is then contoured to obtain the outline of each hair, thereby obtaining a representation of all hairs in the image. The number of contours is then counted to determine the number of hairs in the image.

[0112] In the above embodiment, the efficiency is improved by changing the camera to obtain hair images with different attribute values ​​for the same attribute and then performing segmentation.

[0113] In one embodiment, in combination with the above Figure 5 、 Figure 10 、 Figure 11 as well as Figure 12 As shown, Figure 10 is a flowchart of a hair image processing method in yet another embodiment. Figure 11 is a schematic diagram of a natural light image in one embodiment; Figure 12 is a schematic diagram of a depth image in an embodiment.

[0114] In this embodiment, it is assumed that different cameras are respectively a camera for collecting depth information and a natural light camera. In other embodiments, other different cameras may also be used, and the details can be referred to above. The camera for collecting depth information may be as follows: Figure 5 The image shown is obtained using a natural light camera and a monocular structured light camera, so that natural light images and depth images can be captured at the same time.

[0115] The camera for collecting depth information and the natural light camera are activated by the processing module 100. The doctor moves the camera to the designated area, and the processing module 100 controls the light source module to turn on.

[0116] The camera that captures depth information captures a depth image, while the natural light camera captures a natural light image. Processing module 100 maps the depth image onto the plane of the natural light image using a pre-calibrated projection relationship to obtain a processed image. Processing module 100 then subtracts the pixel matrices of the processed image from the natural light image to obtain a difference image matrix. Based on the difference image matrix, a difference image is generated. Threshold segmentation is then performed on the difference image to identify the hair and scalp regions. The hair regions are then contoured to obtain the outline of each hair, thereby obtaining a representation of all hairs in the image. The number of contours is then counted to determine the number of hairs in the image.

[0117] Specifically, in one embodiment, the difference between hair images corresponding to different attribute values ​​under the same attribute is calculated to obtain a difference image, including: obtaining the projection relationship of different cameras obtained in advance; projecting the depth hair image onto the image plane of the two-dimensional hair image according to the projection relationship to obtain the image to be processed; calculating the difference between the image to be processed and the two-dimensional hair image to obtain the difference image.

[0118] The depth-collecting camera and the natural light camera capture hair. The two images capture different areas due to different perspectives. This difference in perspective is addressed by combining the depth image generated by the depth-collecting camera with the intrinsic matrix between the two cameras. The depth image is calculated to create a point cloud. This point cloud is multiplied by the intrinsic matrix, which is then multiplied by the natural light camera's intrinsic parameter matrix to produce an image with the same perspective as the natural light image.

[0119] Specifically, processing module 100 knows the intrinsic parameter matrix, or the projection relationship between the two cameras, which includes parameters u0, v0, dx, dy, and f. u0 and v0 represent the actual position of the optical center in pixels, dx represents the width of the pixel in the x-direction in millimeters, dy represents the width of the pixel in the y-direction in millimeters, and f represents the focal length. Processing module 100 also knows the fundamental matrix between the two cameras.

[0120] The processing module 100 first converts the depth image into point cloud data, which can be specifically performed using the following formula:

[0121] x pc =Depth×(u-u0)×dx / f

[0122] y pc =Depth×(v-v0)×dy / f

[0123] Z pc =Depth

[0124] Among them, Depth represents the depth image, u and v represent the coordinates of a pixel in the depth image, and x pc 、y pc and z pc Coordinates used to represent point cloud data.

[0125] The processing module 100 converts the point cloud data into an image to be processed, which can be specifically done using the following formula:

[0126] Image=PointCloud×Fundamental Matrix×Intrinsic Matrix

[0127] Among them, Image is the image to be processed, and Intrinsic Matrix is ​​the intrinsic parameter matrix, specifically:

[0128]

[0129] Finally, processing module 100 subtracts the pixel matrices of the image to be processed from the natural light image to obtain a difference image matrix. Based on the difference image matrix, a difference image is generated. This difference image is then thresholded to identify the hair and scalp regions. The hair regions are then contoured to obtain the outline of each hair, thereby obtaining a representation of all hairs in the image. The number of contours is then counted to determine the number of hairs in the image.

[0130] In the above embodiment, the efficiency is improved by changing the camera to obtain hair images with different attribute values ​​for the same attribute and then performing segmentation.

[0131] In one embodiment, performing hair recognition on the difference image to obtain the hair recognition result includes: performing threshold segmentation on the difference image to obtain the hair area; and performing contour extraction on the hair area to obtain the hair recognition result.

[0132] Specifically, combined Figure 13 and Figure 14 As shown, Figure 13 FIG. 1 is a schematic diagram of the result of performing threshold segmentation on the difference image in one embodiment. Figure 14 This is a schematic diagram of an embodiment after contour extraction. The processing module 100 calculates the difference between hair images with different attribute values ​​to generate a difference image, and then performs threshold segmentation on the difference image. This is because black hair absorbs light, resulting in relatively small grayscale value variations in images under different light sources. However, the grayscale value of the scalp region varies significantly under different light sources. Therefore, the value of each pixel in the difference image is compared with a threshold. If the value is less than or equal to the threshold, the pixel is segmented as a hair region; otherwise, it is segmented as background. For example, hair segmentation is accomplished by labeling locations with values ​​close to 0 as hair, while areas with significant variations are considered background.

[0133] In addition, after the hair region is segmented, the contour of the hair region is extracted. The contour extraction method can adopt any existing method, which will not be described here. Figure 14 , which is the image after contour extraction, the processing module 100 can calculate the number of hairs by counting the number of contours.

[0134] In the above embodiment, the hair is segmented by utilizing the property of black hair that it can absorb all light, thereby increasing the segmentation speed.

[0135] In one embodiment, when a binocular camera is used instead of a monocular camera, the binocular camera can also provide hair thickness and / or length information based on its ability to capture 3D image information. The processing module 100 can reproject the identified hair contours into the depth image, thereby determining the hair region in the depth image, and then calculate the hair thickness and / or hair length based on the data in the depth image. In other embodiments, the processing module 100 can determine a conversion relationship between pixel units and length units, such as millimeters, based on the mapping relationship between the depth image and the two-dimensional image. In this way, the pixel length of the short side of the contour in the difference image can be read, and the hair thickness in millimeters can be obtained based on this conversion relationship. Similarly, the calculation of hair length is similar and will not be further described here. In one embodiment, hair thickness and / or hair length can be identified by statistical values ​​of hair thickness and / or hair length in the hair region, such as the average value of hair thickness and / or hair length, and this average value is output for the doctor's reference.

[0136] Specifically, combined Figure 7 As shown, processing module 100 records the contour information of all hairs and stereo information (when using a binocular camera) and displays the results on display 500. The information bar displayed shows the number of hairs, average thickness, and average length. When you move the mouse over a hair, the thickness and length of the current hair are displayed. The thickness of the hair is the width of the contour, and the length of the contour is the length of the hair. The actual width and length are calculated by converting the image into a stereo image.

[0137] In the above embodiment, the black color of hair reduces the influence of ambient light on hair segmentation and counting. By improving the calculation speed of the hair segmentation method, the speed of automatic hair extraction is increased, and the operation time is reduced. By improving the anti-interference performance of hair segmentation, the accuracy of hair positioning and hair extraction is improved. The computational complexity of hair segmentation is reduced, and the number of hairs greater than N is increased. 2 The computational complexity is reduced to 2N, reducing the heat generated by the machine during calculations. This increases the algorithm's universal applicability and avoids significant differences in segmentation results for different ethnic groups. It also opens up the possibility of using high-resolution cameras for hair segmentation, as the computation time does not increase significantly with the increase in resolution. By improving the efficiency and anti-interference performance of hair segmentation, it can quickly count the number of hairs in an image. When using a binocular camera, it can also provide the thickness of large areas of hair in an image, assisting doctors in diagnosis.

[0138] It should be understood that, although the steps in the flowcharts of the above embodiments are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the flowcharts of the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily to be performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0139] Based on the same inventive concept, embodiments of the present application further provide a hair image processing device for implementing the aforementioned hair image processing method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more hair image processing device embodiments provided below can be found in the aforementioned limitations of the hair image processing method and will not be further elaborated here.

[0140] In one embodiment, a hair image processing device is provided, comprising: a hair image acquisition module, a difference image calculation module, and a recognition module, wherein:

[0141] The hair image acquisition module is used to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0142] The difference image calculation module is used to calculate the difference of hair images corresponding to different attribute values ​​to obtain a difference image.

[0143] The recognition module is used to perform hair recognition on the difference image to obtain a hair recognition result.

[0144] In one embodiment, the hair image acquisition module is used to acquire hair images by at least one of the following methods: photographing the same hair area under different light sources using a camera to obtain hair images corresponding to different attribute values ​​under the same attribute; or photographing the same hair area using different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0145] In one embodiment, the hair image acquisition module is used to acquire hair images by at least one of the following methods: changing the attribute value of the light source, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute; or turning on different light sources in sequence, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute.

[0146] In one embodiment, the hair image acquisition module is configured to change the property value of the light source by at least one of the following methods: changing the property value output by the light source, or changing the property value of the light source through a filter.

[0147] In one embodiment, the hair image acquisition module is used to acquire hair images by at least one of the following methods: photographing the same hair area through different cameras to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute; or photographing the same hair area through different cameras to obtain two-dimensional hair images and depth hair images respectively, and the two-dimensional hair image and the depth hair image have the same attributes but different attribute values.

[0148] In one embodiment, the hair image acquisition module is used to acquire hair images by at least one of the following methods: using different cameras to capture images formed by the reflector 600 after reflection in different directions, so as to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute, and the reflector 600 is used to reflect light passing through the same hair area.

[0149] In one embodiment, the difference image calculation module includes:

[0150] The first projection relationship acquisition unit is used to acquire the projection relationships of different cameras obtained in advance.

[0151] The first projection unit is used to project the depth hair image onto the image plane of the two-dimensional hair image according to the projection relationship to obtain an image to be processed.

[0152] The first difference image calculation unit is used to calculate the difference between the image to be processed and the two-dimensional hair image to obtain a difference image.

[0153] In one embodiment, the difference image calculation module includes:

[0154] a second projection relationship acquisition unit, configured to acquire a projection relationship between two cameras in a pre-calibrated binocular camera, wherein the binocular camera includes a monocular camera and a natural light camera;

[0155] a second projection unit, configured to project the hair image captured by one of the cameras onto an image plane of the hair image captured by the other camera according to the projection relationship, to obtain an image to be processed; or

[0156] The second difference image calculation unit is used to calculate the difference between the image to be processed and the hair image captured by another camera to obtain a difference image. In one embodiment, the recognition module includes:

[0157] The segmentation unit is used to perform threshold segmentation on the difference image to obtain the hair area.

[0158] The extraction unit is used to extract the contour of the hair area to obtain the hair recognition result.

[0159] In one embodiment, the apparatus further comprises:

[0160] The statistical module is used to obtain hair attributes based on the hair recognition results, where the hair attributes include at least one of the number of hairs, the thickness of hairs, and the length of hairs.

[0161] Each module in the hair image processing device described above may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0162] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 15 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a hair image processing method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0163] Those skilled in the art will understand that Figure 15 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0164] In one embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0165] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0166] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0167] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0168] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0169] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A hair image processing method, characterized in that: The method comprises: Acquire collected hair images corresponding to different attribute values ​​under the same attribute, wherein the attribute includes color and / or intensity; Calculating differences between hair images corresponding to different attribute values ​​to obtain difference images, including: calculating differences in pixel values ​​at corresponding positions of the hair images to obtain difference images; Performing hair recognition on the difference image to obtain a hair recognition result.

2. The method according to claim 1, characterized in that The acquiring of the collected hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following: Using a camera to photograph the same hair area under different light sources to obtain hair images corresponding to different attribute values ​​under the same attribute; or The same hair area is photographed by different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute.

3. The method according to claim 2, characterized in that The method of photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following: Changing the property value of the light source, photographing the same hair area under different light sources with a camera to obtain hair images corresponding to different property values ​​under the same property; or Turn on different light sources in sequence, and use a camera to shoot the same hair area under different light sources to obtain hair images corresponding to different attribute values ​​under the same attribute.

4. The method according to claim 3, characterized in that Changing the property value of the light source includes: The property value of the light source output is changed, or the property value of the light source is changed through a filter.

5. The method according to claim 2, characterized in that The method of photographing the same hair area with different cameras to obtain hair images corresponding to different attribute values ​​under the same attribute includes at least one of the following: Using different cameras to photograph the same hair area to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute; or The same hair area is photographed by different cameras to obtain a two-dimensional hair image and a depth hair image respectively, and the two-dimensional hair image and the depth hair image have the same attributes but different attribute values.

6. The method according to claim 5, characterized in that The method of photographing the same hair area with different cameras to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute includes: Different cameras are used to capture images formed by the reflector after reflection in different directions to obtain two-dimensional hair images or depth hair images corresponding to different attribute values ​​under the same attribute. The reflector is used to reflect light passing through the same hair area.

7. The method according to claim 5, characterized in that The calculating the difference of the hair images corresponding to different attribute values ​​under the same attribute to obtain the difference image includes: Obtaining pre-calibrated projection relationships of different cameras; Projecting the depth hair image onto the image plane of the two-dimensional hair image according to the projection relationship to obtain an image to be processed; The difference between the image to be processed and the two-dimensional hair image is calculated to obtain a difference image.

8. The method according to claim 5, characterized in that Calculate the difference in hair images corresponding to different attribute values ​​under the same attribute to obtain a difference image, including: Obtaining a projection relationship between two cameras in a pre-calibrated binocular camera, wherein the binocular camera includes a monocular camera and a natural light camera; Projecting the hair image captured by one of the cameras onto the image plane of the hair image captured by the other camera according to the projection relationship to obtain an image to be processed; or The difference between the image to be processed and the hair image captured by another camera is calculated to obtain a difference image.

9. The method according to any one of claims 1 to 8, characterized in that The performing hair recognition on the difference image to obtain a hair recognition result includes: performing threshold segmentation on the difference image to obtain a hair region; Contour extraction is performed on the hair area to obtain a hair recognition result.

10. The method according to claim 9, characterized in that After extracting the contour of the hair area to obtain the hair recognition result, the method includes: Hair attributes are obtained based on the hair recognition results, where the hair attributes include at least one of hair quantity, hair thickness, and hair length.

11. A hair recognition system, characterized in that: The hair recognition system comprises: A light source module, configured to provide a light source for irradiating the hair; A camera module is used to collect hair images corresponding to different attribute values ​​under the same attribute; A processing module, configured to process the collected hair images corresponding to different attribute values ​​under the same attribute according to the hair image processing method according to any one of claims 1 to 10 to obtain a hair recognition result.

12. The system according to claim 11, wherein: The light source module includes a light source with adjustable properties, or the light source module includes an adjustable filter.

13. The system according to claim 11, wherein: The camera module includes different types of two-dimensional cameras, or the camera module includes different types of cameras for collecting depth information; or the camera module includes a camera for collecting depth information and a two-dimensional camera.

14. The system according to claim 13, wherein: The camera module further includes a reflector, which is used to reflect the light from the light source module toward different cameras.

15. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 10 are implemented.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 10 are implemented.

17. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 10 are implemented.

Images