TWI696193B - Body shape prediction system - Google Patents

Abstract

The invention is a body shape change prediction system, which uses a body shape three-dimensional scanning device to collect body shape data of different parts of an individual's body, a physiological data collection unit to collect individual physiological data, the collected body shape data and physiology The data is processed by an artificial intelligence computing host to generate predicted body shape data and physiological data after body shape changes in the future, and converted into a digital three-dimensional portrait model and presented on a display interface for personal viewing reference.

Description

Body shape prediction system

The invention relates to a technology for predicting changes in human body shape, in particular to a prediction system for predicting changes in body shape and appearance according to human physiological characteristic parameters.

Whether it is for health or aesthetic considerations, having an ideal posture is one of the goals pursued by most people. Studies have shown that being overweight or too thin may become the culprit of many chronic diseases. For example, the risk of diabetes, metabolic syndrome, cardiovascular disease, etc. when overweight is higher than the normal number of people; on the other hand, underweight It is easy to cause fatigue, depression, muscle weakness, and may even reduce the body's immune ability. Therefore, many people try to control their posture effectively through various channels.

For example, through continuous exercise, fitness, professional nutrition management and other programs, the purpose of reducing body fat and increasing muscles can be achieved, and the effect of sculpting can be achieved. The aforementioned various types of body management are gradual changes, so the parties need to continue to take action to see the results.

However, when the body shape management plan was originally formulated, the plan may only be a desired value (such as weight and body fat rate), and the parties could not know their body shape changes and final appearance during the planning process. Therefore, it is not easy for the parties to generate motivation for positive action, and most plans may be abandoned halfway.

The main purpose of the present invention is to provide a system for predicting changes in body shape, which predicts changes in individual posture based on the continuously collected personal physiological characteristic parameters.

In order to achieve the aforementioned objective, the "body change trend prediction system" of the present invention mainly includes: A body shape three-dimensional scanning device, used to scan the body shape to obtain body shape data of different parts of the body; A physiological data collection unit, used to measure personal physiological data; An artificial intelligence computing host, including: A database, which stores body shape data and physiological data measured by different individuals using the body three-dimensional scanning device and the physiological data collection unit; A learning engine accesses the body shape data and physiological data recorded in the database to learn and generate predicted body shape data based on the measured body shape data and the physiological data of the individual; A display interface outputs predicted figure data generated by the learning engine, wherein the display interface displays a digital three-dimensional portrait model based on the predicted figure data.

With the above-mentioned body shape prediction system of the present invention, it is possible to estimate the amount of possible body shape changes in the future by collecting personal body shape data and physiological data, for example, it is used to intentionally change oneself through sports fitness, diet adjustment or other methods The users of the figure can know the possible future appearance as the goal at the early stage of planning.

Furthermore, the present invention can continuously update the changes of the personal shape by sequentially recording the changes of the same personal shape data and physiological data to provide more real-time reference information.

Please refer to FIG. 1, the present invention is a body shape change trend prediction system, which includes a body three-dimensional scanning device 10, a physiological data collection unit 20 and an artificial intelligence computing host 30. For parties who plan to change their body shape, for example, if the party wants to achieve weight loss through surgical weight loss (such as gastric bypass surgery), diet control or exercise and other progressive actions, the present invention can be measured by the relevant data of the client. It predicts the posture change trend in the future and provides the future change trend of possible appearance as a reference.

The body shape three-dimensional scanning device 10 is used to scan a person's body shape to obtain body shape data of different parts of the body, such as face, neck, chest, abdomen, buttocks, right arm, left arm, right thigh, left thigh, etc. According to the body shape data of these different parts, the circumference of each body part can establish a digital three-dimensional body model representing the party.

The physiological data collection unit 20 is used to measure individual's physiological data. The included measurement items include but are not limited to the following parameters: Body composition analysis: Total weight, total water weight, mineral weight, protein weight, body fat weight Obesity analysis: Body Mass Index (BMI), Body Fat Rate (PBF) Physiological index analysis: Blood pressure, blood sugar, heart rate, total cholesterol Muscle analysis by site: Right arm, left arm, torso, right leg, left leg Fat analysis by site: Right arm, left arm, torso, right leg, left leg

The physiological data collection unit 20 may include a fixed measurement instrument or a wearable measurement instrument. In a preferred embodiment, the measurement instrument used adopts a non-invasive measurement technique to measure the physiological data of the human body. Among them, the overall weight of the human body is mainly composed of four types of weight: total water weight, mineral weight, protein weight, and body fat weight. Part-by-part muscle analysis is used to measure the amount or ratio of muscles occupied by various parts of the body. Please refer to Figure 2, which mainly measures the muscle mass of the right arm A, left arm B, trunk C, right leg D, and left leg E; In the same way, site-specific fat analysis is used to measure the amount or ratio of fat occupied by various parts of the body. The above physiological data are closely related to changes in personal body shape data.

The artificial intelligence computing host 30 analyzes the future body shape of the party according to the body shape data of different parts measured by the body shape three-dimensional scanning device 10 and the physiological data obtained by the physiological data collection unit 20.

The artificial intelligence computing host 30 includes a database 31 and a learning engine 32, wherein the database 31 stores body shape data and physiology data of different individuals, and these data not only include the latest body shape data/latest measurement of the last time measured by the same person Physiological data, also including historical figure data/historical physiological data measured in the past; as the number of individuals stored in the database 31 increases, huge amounts of data with different figures and physiological data can be gradually accumulated, making the learning engine The prediction accuracy of 32 has gradually improved.

The learning engine 32 refers to the body shape data and physiological data in the database 31, and learns to generate personal predicted body shape data and predicted physiological data. For example, the user A obtains the latest body shape data/latest physiological data after measuring through the body shape three-dimensional scanning device 10 and the physiological data collection unit 20. The learning engine 32 can find the user from the database 31 A reference person R whose data is closest to each other, based on the body shape change data recorded by the reference person R in the database 31, calculates and outputs the user A in a period of time (such as 3 months, 6 months, etc.) ) After the predicted body shape data and the predicted physiological data, a digital three-dimensional portrait model 50 is established based on the predicted body shape data and the predicted physiological data. The digital three-dimensional portrait model 50 can be further output to a display interface 40 for the user to view and refer to.

Referring to FIG. 3, if the user performs body shape prediction for the first time, the data measured by the user is stored as the original data in the database 31, and the learning engine 32 may further be based on a forward direction set by the user The change parameter or a reverse change parameter determines the change trend of the figure. The forward change parameter represents that the user expects to be thinner to reduce the figure data. When specified as a positive change, the learning engine 32 calculates The predicted figure data is the target value after thinning. On the contrary, the reverse change parameter represents that the user expects to increase the body shape data. When the reverse change is specified, the predicted body shape data calculated by the learning engine 32 is the target value after the increase. When the learning engine 32 considers the forward change parameter or the reverse change parameter, it can refer to the historical figure data/historical physiological data of different individuals recorded in the database 31 to determine the prediction after the forward change or the reverse change data.

As shown in FIG. 4, it is a flow chart of steps for the learning engine 32 to generate the predicted body shape data and the predicted physiological data according to the original data of the parties, including the following steps:

S41: Obtain the original body shape data and original physiological data of the parties. This step is to use the body shape three-dimensional scanning device 10 and the physiological data collection unit 20 to obtain the original body shape data and the original physiological data after the first scan and measurement of the person. The learning engine 32 receives the original body shape data and The original physiological data is used to perform subsequent operations.

S42: According to the sex of the client, data from the sex database 31 is used. Considering that there are obvious differences in the body shapes of men and women, for example, the data of bust and hip circumference are generally significantly different, so the learning engine 32 will select the same gender data from the database 31 as the basis for the subsequent calculation.

S43: Obtain the target value of the person's figure plan. The parties can set the desired body shape target value before carrying out the action of changing their body shape. For example, they hope to obtain the body shape data of each part after successful weight loss as the body shape target value, and use the body shape The shape plan target value is imported into the learning engine 32. In the process of computing the predicted figure data, the learning engine 32 will refer to the target value of the prepared figure plan for calculation. Among them, when the target value of the body shape plan is formulated, it can further determine whether the target value of the body shape plan of the party is within a reasonable range of human safety. If it exceeds the safety range and may cause harm to human health, the learning engine 32 Can provide suggested data for parties' reference.

S44: Screen out the corresponding reference samples in the database. The learning engine 32 draws the corresponding reference samples from the database 31 according to the basic information of the person and the target value of the figure, for example, from the database 31 according to the height, age or other basic information of the person Filter out the data of others corresponding to the basic information as a reference sample. Assuming the height of the person is 172cm, you can find a sample with a height between 171 and 175cm from the database 31, and then further filter out the data of other people of similar age as a reference sample For subsequent learning operations. The learning engine 32 can further select reference samples that conform to this method according to the body shape change method adopted by the person, such as weight loss, diet control, or exercise.

S45: Identify the characteristics of various parts of the body. The learning engine 32 can recognize the circumference values of various body parts such as face, neck, chest, abdomen, buttocks, right arm, left arm, right thigh, and left thigh based on the original figure data acquired in step S41. The surrounding value of each part is regarded as a characteristic value.

S46: Calculate and adjust the feature values of each part of the body to generate the predicted figure data. The learning engine 32 calculates different change trends of the characteristic values of the body parts of the parties in the future based on the reference samples selected from the database 31. In a preferred embodiment, the learning engine 32 will calculate the predicted values of the feature values of each part at different points in the future, such as calculating the first month, the third month, the sixth month, and the ninth month, The possible values for the 12th month are obtained from multiple sets of predicted values at different time periods. When the learning engine 32 calculates the predicted value, one of the methods is to calculate the predicted value of the party according to the rate of change of the corresponding body parts in the reference samples. For example, according to the abdominal circumference data in the filtered reference sample The amount of change over a period of time (such as an average decrease of 2% per month) is used as a benchmark to calculate the forecast data of the party's abdominal circumference at different points in the future. The foregoing example method is only one of them. The learning engine 32 can use other different Learn how to perform calculations.

S47: Calculate and modulate physiological data to generate physiological prediction data. The learning engine 32 calculates the predicted values of the feature values of each part at different time points in the future in the foregoing step S46, and also refers to the data of the reference samples to calculate the physiological data at different time stages from the original physiological data to generate predicted physiological data .

After the user’s first body shape prediction, the user can measure the latest body shape data/newest data through the body shape three-dimensional scanning device 10 and the physiological data collection unit 20 again after a certain time (for example, every 1 to 2 months) The physiological data, the latest data are stored in the database 31. The learning engine 32 compares with the original prediction data to determine the change amount of the body shape data this time and the difference from the target value. Please refer to Figure 5 for the specific process, which includes the following steps:

S51: Obtain the latest body shape data/latest physiological data of the person concerned. This step is the latest data obtained by scanning and measuring the person through the body 3D scanning device 10 and the physiological data collection unit 20.

S52: Compare the latest body shape data with the predicted body shape data to obtain the difference. The learning engine 32 compares the latest body shape data with the previously calculated body shape data. When comparing, the body shape data of the same period will be used as the basis, that is, if the party is the third The latest data obtained from the monthly measurement will also be compared with the forecast data of the third month.

S53: Determine whether the difference is within a set range. The learning engine 32 judges whether the difference is within the setting range according to a preset setting range. If it is, the representative party is currently close to the predicted body shape; if not, the representative party is currently not close to the predicted body shape.

S54: If the difference is within the set range, store the latest body shape data/latest physiological data of the person in the database 31.

S55: If the difference exceeds the set range, the learning engine 32 will re-recognize the feature values of various parts of the body, that is, recognize the various parts of the body such as face, neck, chest, abdomen, buttocks, right arm, left arm, right The latest values of thighs and left thighs. Based on the newly recognized data, the deformation parameters of each part are calculated. The calculation operation includes calculating the deformation ratio of each part of the current body shape, and recalculating the physiological data based on the calculated data of each part.

S56: Store the latest body shape data and the latest physiological data. Because the difference is within the set range, the current body of the representative is not close to the predicted body. The learning engine 32 regards the latest data measured this time as a new set of available data and adds it to the database 31 In order to increase the samples in the database 31 and expand the available information in the database 31.

The body shape change trend prediction system of the present invention can estimate future body shape changes based on personal body shape data and physiological data. The artificial intelligence computing host 30 refers to the huge amount stored in the database 31 based on the measured data Data to generate a predicted figure. In practical applications, physicians, nutritionists, or professionals can use the present invention to predict the posture of users who intend to change their body shape in the future, so that the parties can know their possible future appearance in advance, and improve the sustainability of the parties. Motivation for action; and in the process, by regularly tracking changes in personal body shape data and physiological data, future body shape change data can be updated at any time.

10: Figure 3D scanning device 20: physiological data acquisition unit 30: Artificial intelligence computing host 31: Database 32: Learning engine 40: Display interface 50: digital 3D portrait model

Fig. 1: Block diagram of the architecture of the body shape trend prediction system of the present invention. Figure 2: Schematic diagram of site-specific muscle analysis in the present invention. Figure 3: A schematic diagram of the present invention predicting the forward or reverse change of a person's figure. Fig. 4: Flow chart of the prediction engine of the present invention for generating body shape data. Figure 5: A flowchart of the comparison between the latest data and the original predicted data of the learning engine of the present invention.

10: Figure 3D scanning device

20: physiological data acquisition unit

30: Artificial intelligence computing host

31: Database

32: Learning engine

40: Display interface

50: digital 3D portrait model

Claims (7)

A body shape change prediction system, including: a body shape three-dimensional scanning device for scanning a person's body shape to obtain body shape data of different parts of the body; a physiological data collection unit for measuring the body's physiological data; an artificial The intelligent computing host includes: a database that stores body data and physiological data of different parties measured by the body 3D scanning device and the physiological data collection unit; a learning engine accesses the database The recorded body shape data and physiological data are used to learn to generate predicted body shape data and generate predicted physiological data based on the body shape data measured by the parties and the physiological data; a display interface to output the predicted body shape generated by the learning engine Data, wherein the display interface outputs the predicted physiological data, and displays a digital three-dimensional portrait model based on the predicted body shape data in combination with the predicted physiological data; wherein, the learning engine performs the following steps to generate the predicted body shape data and the predicted physiological data Data: Obtain the original body shape data and original physiological data of the party, wherein the original body shape data is the data obtained by the body shape three-dimensional scanning device for the first scan of the party, and the original physiological data is the physiological data collection unit The data collected by the party for the first time; according to the gender of the party, the body shape data and physiological data of the same gender are taken from the database; the target value of the body shape preset by the party is obtained; based on the basic information of the party and the body Shape plan target value, select the corresponding body shape data and physiological data from the database as a reference sample; identify the surrounding value of each part in the original body shape data; calculate and adjust these surrounding values, To generate the predicted figure data; According to the predicted body shape data, the original physiological data is calculated and modulated to generate the predicted physiological data. The body shape change prediction system according to claim 1, wherein the learning engine generates predicted body shape data and predicted physiological data after the body shape becomes thin according to a set forward change parameter; or according to a set reverse direction The operation of changing parameters generates predicted body shape data and predicted physiological data after the body shape becomes strong. The body shape change prediction system according to claim 1, wherein the predicted body shape data includes multiple sets of data at different time stages; and the predicted physiological data includes multiple sets of data at different time stages. The body shape prediction system according to claim 1, wherein the body shape three-dimensional scanning device generates various parts of the human face, neck, chest, abdomen, kidney, right arm, left arm, right thigh, and left thigh Of the surrounding value. The body shape change prediction system according to claim 1, wherein the physiological data measured by the physiological data measuring unit includes: body composition analysis: total weight, total water weight, mineral weight, protein weight, body fat weight ; Obesity analysis: Body mass index (BMI), body fat rate (PBF); Physiological indicators analysis: blood pressure, blood sugar, heart rate, total cholesterol; analysis of muscle by site: right arm, left arm, trunk, right leg, left leg; Fat analysis by site: right arm, left arm, torso, right leg, left leg. According to the body shape change prediction system of claim 5, the physiological data measurement unit includes a fixed measurement instrument and a wearable physiological measurement instrument. According to the body shape change prediction system described in claim 1, the database stores historical body shape data/historical physiological data measured by the same person through the body three-dimensional scanning device and the physiological data measuring unit.

Images