TWI670047B - Scalp detecting device - Google Patents

Abstract

The invention relates to a scalp detection device, which includes a scalp detection machine, which includes a power supply module, a photographic module electrically connected to the power supply module, and a first wireless transmission unit electrically connected to the power supply module; and An electronic device is electrically connected to the first wireless transmission unit by a second wireless transmission unit. The electronic device also has a scalp identification module and a result output module. The scalp identification module divides the scalp into a plurality of blocks. To identify the image of the block; by this means, the scalp detection device can view the image of the subject's scalp and hair to detect the scalp condition, and the detection result will be stored on the cloud platform for future reference.

Description

Scalp detection equipment

The invention relates to a scalp detection device, in particular to a device capable of looking at the condition of the scalp surface and automatically identifying and detecting the condition of the scalp of the subject. Through deep learning training and identification, the scalp can be detected. The image accurately analyzes whether it is normal or abnormal.

According to the scalp, it belongs to the very sensitive skin in the human body, and there are not a few people who have scalp allergies or prone to dandruff. In addition to seeking medical treatment from a professional doctor, in daily life, many consumers also choose to buy their own hair Care-related products to improve your scalp or hair, or to avoid shampoo products that are more chemically irritating, as the use of these chemicals over time will hurt the hair follicles of our scalp, and the residue will be in the The accumulation of blockage on the scalp causes abnormalities in the scalp, which results in various scalp symptoms or hair loss. The human head is the center of mental activity, and also controls the command post and main sensory organs of the entire body. Modern people are busy and stressed at work. The scalp relief treatment is aimed at chronic stress and tension, which affects the scalp health. For those with scalp problems, they can use the massage technique of balanced pressure relief to restore the unbroken body, body and hair, and show their new vitality.

Today, many hair management and chain hairdressers also offer scalp detection services, and recommend related physiotherapy products and treatments. The current detection methods for scalp conditions are mostly single-point photos of the scalp using detection equipment, and judged by manual interpretation. The condition of the scalp of the test subject; however, the person in charge of interpretation must have professional training to determine the condition of the scalp. Generally, beauty and hairdressing industry operators need to do the education and training of scalp identification before entering the service. Increase the cost of their education and training, but the correctness of human identification after different personnel receive education and training varies from person to person, which often causes distress to testees or consumers and reduces their sense of trust. Therefore, how to provide a detection device that can improve the accuracy of scalp condition interpretation, and even achieve the goal of reducing the money and time cost of extra manpower training required by the industry.

Now, the inventor is considering that the above existing scalp detection instruments still have many missing points in practical implementation, so it is a tireless spirit, supplemented by its rich professional knowledge and years of practical experience to improve it. Based on this, the invention was developed.

The main purpose of the present invention is to provide a scalp detection device, which can automatically identify the skin quality of the subject's head by using the deep learning technology of the device, so as to assist beauty salons or hair management operators to reduce their education and training costs. And enhance consumer satisfaction and trust.

In order to achieve the above-mentioned implementation objective, a scalp detection device of the present invention includes a scalp detection machine, which includes a power supply module, a photographing module electrically connected to the power supply module, and a first wireless connection to the power supply module. A transmission unit; and an electronic device that is electrically connected to the first wireless transmission unit by a second wireless transmission unit, and the electronic device has a scalp identification module and a result output module, and the scalp identification module divides the scalp into a plurality of numbers Block to identify the image of the block.

In one embodiment of the present invention, the power supply module is a secondary battery and is charged by USB.

In one embodiment of the present invention, the camera module further includes a camera lens and a fill light.

In one embodiment of the present invention, the photographic lens is made of a Charge-coupled Device (CCD) or a Complementary Metal-Oxide-Semiconductor (CMOS).

In one embodiment of the present invention, the scalp recognition module uses deep learning technology.

In one embodiment of the present invention, the scalp recognition module has a training unit and a recognition unit.

In one embodiment of the present invention, the training unit has an image database and a hypothetical function library. Through the hypothetical function library, functions suitable for the photos provided by the image database are found.

In one embodiment of the present invention, the scalp recognition module recognizes the colors of a plurality of blocks, and combines the features according to the color distribution of the plurality of blocks to identify whether the scalp has a bacterial infection, an allergy, or folliculitis, Dandruff, excess oil and hair loss.

In one embodiment of the present invention, the electronic device may be one of a tablet computer, a smart phone, a desktop computer, or a portable computer, and the electronic device is further electrically connected to a cloud platform.

In one embodiment of the present invention, the cloud platform further includes a user database.

(1) ‧‧‧Scalp Detector

(11) ‧‧‧Power supply module

(12) ‧‧‧Photographic Module

(121) ‧‧‧Photographic lens

(122) ‧‧‧ Fill Light

(13) ‧‧‧The first wireless transmission unit

(2) ‧‧‧Electronic equipment

(21) ‧‧‧Second wireless transmission unit

(22) ‧‧‧Scalp Identification Module

(221) ‧‧‧Training Unit

(2211) ‧‧‧Image Database

(2212) ‧‧‧ Hypothetical Function Library

(222) ‧‧‧Identification Unit

(23) ‧‧‧Result output module

(3) ‧‧‧Cloud Platform

(31) ‧‧‧User Database

FIG. 1 is a block diagram of a device architecture according to a preferred embodiment of the present invention.

Second figure: Appearance view of the scalp detector of the preferred embodiment of the present invention (1).

Third figure: Appearance view of the scalp detector of the preferred embodiment of the present invention (2).

FIG. 4 is a schematic diagram of scalp feature extraction according to a preferred embodiment of the present invention.

Fifth Figure: Comparison of identification features of the preferred embodiment of the present invention.

The purpose of the present invention and its structural and functional advantages will be explained based on the structure shown in the following drawings, in conjunction with specific embodiments, so that the reviewing committee can have a deeper and more specific understanding of the present invention.

Please refer to the first to third figures. A scalp detection device according to the present invention includes a scalp detector (1), which includes a power supply module (11) and a photographic module electrically connected to the power supply module (11). The group (12) is electrically connected to a first wireless transmission unit (13) of a power supply module (11). The power supply module (11) may be, for example, a secondary battery. When the secondary battery is exhausted, It can be charged through USB. The camera module (12) further has a camera lens (121) and a fill light (122). The camera lens (121) is made of CCD or CMOS as a light sensor. The scalp detector ( The physical appearance of 1) is shown in the second and third figures; and an electronic device (2) may be, for example, a tablet computer, a smart phone, a desktop computer, or a portable computer, which is electrically connected to a The cloud platform (3) is electrically connected to the first wireless transmission unit (13) by a second wireless transmission unit (21). The electronic device (2) has a scalp identification module (22) and a result output module ( 23). The cloud platform (3) includes a user database (31). The scalp identification module (22) divides the scalp into multiple blocks to identify the color characteristics of the image of the block, including identifying the color and distribution within the block, and can detect whether the scalp has a bacterial infection, allergy, or folliculitis , Dandruff, excess oil, and hair fall, the detection results will be generated by the result output module (23). Because the scalp identification module (22) uses deep learning technology, the scalp identification module (22) will have a A training unit (221) and a recognition unit (222), where the training unit (221) includes an image database (2211) and a hypothetical function library (2212), which are found through the hypothetical function database (2212) and the image database (2211) The photo fit function provided.

In addition, through the following specific examples, the scope of the present invention can be further proved, but it is not intended to limit the scope of the present invention in any form.

Please continue to refer to the first to third figures, the scalp detection device of the present invention is divided into scalp The detector (1) and the electronic device (2) identify the condition of the subject's scalp through the scalp identification module (22) of the electronic device (2), and the scalp identification module (22) is implemented by deep learning technology. The learning method is to first establish a network architecture capable of selecting a hypothesis function, and then compose a scalp recognition module according to the training program and CNN architecture (22). The present invention uses a Convolutional Neural Network (CNN). This is mainly to use a computer to imitate the recognition method of the human brain.The main architectures are a Convolutional Layer, a Rectified Linear Units Layer (ReLU layer), a Pooling Layer, and a Normalization Layer. , Fully connected layer, Logistic regression layer (Softmax Layer) and loss function layer (Loss Layer); the objective function is selected according to the form of training, and multiple sets of functions are used to find the best set of The combination of parameters allows the prediction result to be narrowed down to the target. Among them, the training process is performed by the training unit (221) of the scalp recognition module (22) for recognition training. Image data must be prepared before training (2211) and hypothetical function library (2212), so that the training unit (221) finds a function that is most suitable for the photos provided by the image database (2211) from a large number of functions in the hypothetical function library (2212). Accuracy, the inventor in the image database (2211) prepared 30,000 various scalp photos as training image data. During the training process, there are input and output ends. The input end is the photos of the image database (2211). The output end is the answer for photo recognition. According to the present invention, the answer is bacterial infection, allergy or folliculitis, dandruff, excess fat and hair loss. During training, the characteristics of various scalp conditions are used as multiple training parameters. Training units (221) It will extract whether there are the same range of the characteristics of the training parameters in the plurality of blocks in the photo, multiply the training parameters by a weights, add all the products and add an offset value (bias ), And then get the answer at the output through the Activation Function, which means finding the most suitable function in the hypothetical function library (2212), so that the scalp recognition module (22) can distinguish each The condition of the scalp, i.e., identification of training as shown in FIG. Fourth, the scalp divided into a plurality of blocks to identify blocks Features and find suitable functions.

In actual operation, first press the power switch of the scalp detector (1) to start the power supply module (11). When the LED of the housing is constantly on, it means that the scalp detector (1) has been started, and then the first wireless transmission is turned on. The unit (13); the electronic device (2) is connected to the scalp detector (1) through a second wireless transmission unit (21), and the first wireless transmission unit (13) and the second wireless transmission unit (21) may be, for example, Use WIFI or Bluetooth wireless technology to transmit and receive real-time images of the scalp output by the camera module (12), and also capture the current image as a photo.

During use, the camera lens (121) of the camera module (12) of the scalp detector (1) is pointed at the scalp of the subject, and the scalp image will be transmitted to the second wireless transmission through the first wireless transmission unit (13). The unit (21), the screen of the electronic device (2) will display the picture of the scalp image, and the image captured by the photographic lens (121) is magnified 200 times, so you can see the details on the scalp. Press the fill light (122) button of the camera module (12) to turn on the supplemental light source of the LED to illuminate the detection part of the scalp. The most common conditions of the scalp are bacterial infections, allergies or folliculitis, dandruff, and oil. Excess and hair loss.

During the test, you can refer to the fifth chart, which is divided into four types: bacterial infection, allergy or folliculitis, 2. dandruff, 3. excess fat, and 4. hair loss. The scalp identification module (22) will Identify the scalp characteristics based on these four conditions. Taking the identification of hair loss as an example, because the scalp condition belongs to hair loss, the entire scalp will show a large area of skin color, and there are fewer black blocks. Although the color of the skin may be affected by light, it may be It will have similar characteristics to excess oil, but its color features overlap too little, so the identification unit (222) will not judge the scalp to have the problem of excess oil. Finally, the identification unit (222) of the scalp identification module (22) will According to the results of the previous training unit (221), it can identify the part that has a lot of overlap with the features of hair loss. The result output module (23) of the electronic device (2) will also show the highest score of hair loss features, which represents this time. The result of a scalp test was hair loss.

It can be seen from the foregoing that each scalp condition has different characteristics, and the color displayed on the scalp will also be different. Thus, the detection mode of the identification unit (222) has different identification detection directions according to different scalp conditions; 1. When the scalp is infected with bacteria or the scalp has an allergic reaction, it will cause red itch on the scalp, and it will appear as a large number of small red dots. When hair folliculitis occurs, the root of the hair on the scalp will be large. Block blood red areas, so when identifying bacterial infections, allergies or folliculitis, mainly look for the red block area existing in the skin color area; 2. Dandruff is mainly characterized by white blocks of varying numbers appearing on the scalp , And there is a great chance that it will accompany the appearance of hair. Therefore, when identifying the characteristics of dandruff, it is mainly to find the area of white blocks around the hair area existing in the skin color area; 3. The characteristic of excess oil is the excess oil of the scalp. The large number of white areas caused by the ambient light source or the light source of the supplementary light (122) makes the white of the block a high-brightness white. Therefore, when identifying excess oil, it is basically based on the proportion of the area occupied by high-brightness white blocks. 4. The most important feature of the phenomenon is that the black areas representing the hair are very rare, and there will be large areas of skin color. When the phenomenon is found, it is basically to detect the black hair blocks, and to identify whether there are a large number of skin areas; it can be seen that the scalp recognition module (22) judges the scalp according to the color characteristics of multiple blocks on the scalp. Is it normal?

Finally, the test result will be displayed on the electronic device (2) by the result output module (23). The tester can upload the test result of each test subject to the cloud platform (3), and the user database (31) In the establishment of the subject's exclusive scalp test records, beauty and hairdressers who provide scalp physiotherapy can track consumers' efficacy after using scalp care treatments.

As can be seen from the foregoing implementation description, compared with the prior art, the present invention has the following advantages:

1. The scalp detection device of the present invention uses a scalp detection machine and electronic equipment to be electrically connected to each other, and uses a camera module of the scalp detection machine to take an image of the scalp of the subject, and then The real-time image is transmitted to the electronic device, and the displayed image is divided into a plurality of blocks through the scalp recognition module to identify the condition of each block of the scalp. The invention systematically recognizes the results and reduces artificial recognition. Errors, and improve consumer trust with objective analysis results, and can reduce the cost of money and time for operators to require additional manpower training.

2. The scalp recognition module of the scalp detection device of the present invention trains how to recognize the condition of the scalp through deep learning technology, and allows a training unit to learn various scalp conditions with a large number of photos, so that the recognition unit actually recognizes the scalp image of the subject, Can identify the color of multiple blocks of the scalp to detect whether the features around the hair and on the scalp have bacterial infections, allergies or folliculitis, dandruff, excess oil and hair loss. This detection method has high accuracy and can be directly Let the inspector know the scalp condition of the subject, and there will be no different detection results due to different inspectors.

In summary, the scalp detection device of the present invention can indeed achieve the expected use effect through the above-disclosed embodiments, and the present invention has not been disclosed before the application, and it has fully met the requirements and requirements of the Patent Law. . I filed an application for an invention patent in accordance with the law, and I urge you to examine it and grant the patent.

However, the illustrations and descriptions disclosed above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Anyone who is familiar with the technology, according to the characteristic scope of the present invention, makes other Equivalent changes or modifications should be regarded as not departing from the design scope of the present invention.

Claims (6)

A scalp detection device comprising: a scalp detection machine including a power supply module, a photographic module electrically connected to the power supply module, and a first wireless transmission unit electrically connected to the power supply module; and An electronic device is electrically connected to the first wireless transmission unit by a second wireless transmission unit. The electronic device further has a scalp recognition module and a result output module. The scalp recognition module has a training unit and a A recognition unit that uses deep learning technology to divide the scalp into a plurality of blocks to identify the color of the images of the plurality of blocks, and to identify whether the scalp has a bacterial infection, allergy or Folliculitis, dandruff, excess oil, and hair loss. The training unit has an image database and a hypothetical function library. Through the hypothetical function library, functions suitable for the photos provided by the image database are found. The scalp detection device according to item 1 of the scope of patent application, wherein the power supply module is a secondary battery, and the secondary battery is charged by USB. The scalp detection device according to item 1 of the patent application scope, wherein the camera module further has a camera lens and a fill light. The scalp detection device according to item 3 of the scope of the patent application, wherein the photographic lens is made of a charge-coupled device (CCD) or a complementary metal-oxide semiconductor (CMOS). For example, the scalp detection device described in item 1 of the patent application scope, wherein the electronic device is one of a tablet computer, a smart phone, a desktop computer, or a portable computer, and the electronic device is further electrically connected to a cloud platform. According to the scalp detection device described in item 5 of the patent application scope, the cloud platform further includes a user database.