CN110051353A - A kind of body fat assessment system and method based on depth learning technology - Google Patents
A kind of body fat assessment system and method based on depth learning technology Download PDFInfo
- Publication number
- CN110051353A CN110051353A CN201910365560.XA CN201910365560A CN110051353A CN 110051353 A CN110051353 A CN 110051353A CN 201910365560 A CN201910365560 A CN 201910365560A CN 110051353 A CN110051353 A CN 110051353A
- Authority
- CN
- China
- Prior art keywords
- body fat
- data
- deep learning
- client terminal
- cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves
- A61B5/053—Measuring electrical impedance or conductance of a portion of the body
- A61B5/0537—Measuring body composition by impedance, e.g. tissue hydration or fat content
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/48—Other medical applications
- A61B5/4869—Determining body composition
- A61B5/4872—Body fat
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
Abstract
The invention discloses a kind of body fat assessment system and method based on depth learning technology, it can be according to human body stereo dynamic image, in conjunction with the measurement result of intelligent Human fat balance, it and include the cloud experience database of a large amount of empirical datas, human body body fat rate, muscle rate, protein content, bone amount and moisture content are made and are more in line with actual assessment, wherein cloud experience database has the function of deep learning, can guarantee that internal empirical data is most accurately data always.
Description
Technical field
The invention belongs to artificial intelligence Visual identification technology fields, specifically, being to be related to one kind based on deep learning skill
The body fat assessment system and method for art.
Background technique
Currently, various intelligence Human fat balances are very popular, but these scales are substantially based on the measurement of BIA biologic resistance
Method measurement skin resistance etc. is as according to being estimated, since the degree of drying and temperature of skin vary with each individual, because of Shi Eryi, because
This error is larger.
Summary of the invention
It is commented the object of the invention is that providing a kind of body fat based on deep learning technology to solve the above-mentioned problems
Estimate system and method, the projects such as human body body fat rate, muscle rate, protein content, bone amount and moisture can be made and more accord with
Close actual assessment.
The present invention through the following technical solutions to achieve the above objectives:
A kind of body fat assessment system based on deep learning technology, including cloud experience database, client terminal machine and have
The intelligent Human fat balance of skin temperature and humidity surveying function establishes data between the cloud experience database and the client terminal machine
Interaction channel establishes data exchange channels, the cloud empirical data between the client terminal machine and the intelligent Human fat balance
Record has experience data in library, and the empirical data is classified according to the gender of people, age, figure and ethnic group, and including
Body fat rate, muscle rate, protein content, bone amount, moisture content and anatomical data；
The cloud experience database has the function of deep learning, can constantly acquire the new number different from original empirical data
According to, and by original empirical data in conjunction with new data after re-start classification.
Preferably, the intelligence Human fat balance is the Human fat balance based on BIA biologic resistance mensuration.
Preferably, the conductive film of the intelligence Human fat balance is ITO conductive film, it is embedded with thermocouple, can measure and be connect
The skin temperature of touching, the temperature are used to adjust the resistance value of sampling.
Preferably, being equipped with moisture probe inside the intelligence Human fat balance, ambient humidity can be measured, the ambient humidity
For adjusting the resistance value of sampling.
Preferably, the client terminal machine is smart phone or PC machine.
A kind of body fat appraisal procedure based on deep learning technology is assessed using the above-mentioned body fat based on deep learning technology
System carries out body fat assessment to user, comprising the following steps:
1) for subscriber station on intelligent Human fat balance, intelligent Human fat balance measures the body fat of user, and by measured data
Be transferred to client terminal machine, which is uploaded to cloud experience database by client terminal machine, and in the experience database of cloud
Original empirical data compare；
2) user makes a series of dynamic actions, and client terminal machine carries out the record of figure dynamic 3-D video to user, and will
The figure dynamic 3-D video data are uploaded to cloud experience database, and with original empirical data in the experience database of cloud
It compares；
3) client terminal machine is obtained according to step 1) and step 2 data comparison as a result, obtain the body fat rate of user, muscle rate,
Protein content, bone amount and moisture content.
Preferably, a series of dynamic actions for making of user include turn left 90 degree, rear turnback, turn right 45 degree, bend over
90 degree, arm it is flattened.
Compared with prior art, the invention has the following advantages:
Can be according to human body stereo dynamic image, in conjunction with the measurement result of intelligent Human fat balance, and include a large amount of empirical datas
Cloud experience database is made to human body body fat rate, muscle rate, protein content, bone amount and moisture content and is more in line with reality
The assessment on border, wherein cloud experience database has the function of deep learning, can guarantee that internal empirical data is most smart always
Quasi- data.
Detailed description of the invention
Fig. 1 is the structural schematic diagram of the body fat assessment system of the present invention based on deep learning technology；
In above-mentioned attached drawing, the corresponding component names of appended drawing reference are as follows:
1- user, the cloud 2- experience database, 3- client terminal machine, 4- intelligence Human fat balance.
Specific embodiment
The present invention will be further explained below with reference to the attached drawings:
As shown in connection with fig. 1, the body fat assessment system of the present invention based on deep learning technology, including cloud experience database 2,
Client terminal machine 3 and intelligent Human fat balance 4 with skin temperature and humidity surveying function, cloud experience database 2 and client terminal
Data exchange channels are established between machine 3, establish data exchange channels, cloud experience between client terminal machine 3 and intelligent Human fat balance 4
Record has experience data in database 2, and empirical data is classified according to the gender of people, age, figure and ethnic group, and including
Body fat rate, muscle rate, protein content, bone amount, moisture content and anatomical data；
Cloud experience database 2 has the function of deep learning, can constantly acquire the new data different from original empirical data, and
By original empirical data in conjunction with new data after re-start classification, constantly adjustment updates in this way, can make whole system
It is higher and higher that body fat assesses accuracy.
In the present system, intelligent Human fat balance 4 is the Human fat balance based on BIA biologic resistance mensuration, and Human fat balance passes through electrode
Piece issues weak current, forms a closed loop with human body, and by muscle easy conductive, the nonconducting principle of fat obtains an electricity
Resistance numerical value is simultaneously uploaded to client terminal machine 3.
In the present system, the conductive film of intelligent Human fat balance 4 is ITO conductive film, logical using one foot of micro current human body
Buttocks is crossed to the resistance of another foot, therefore there cannot be any contact when use between two legs and bipod, ITO conductive film is embedded with warm
Galvanic couple can measure contacted skin temperature, which is used to adjust the resistance value of sampling.
In the present system, moisture probe is installed inside intelligent Human fat balance 4, ambient humidity can be measured, the ambient humidity
For adjusting the resistance value of sampling.
In the present system, client terminal machine 3 is smart phone or PC machine, in order to enable client terminal machine 3 has accordingly
Function, such as data transmission saves, analysis, mountable corresponding APP on client terminal machine 3.
In fact, the figure of people, fat-muscle distribution situation are different, and with ethnic group and age size variation, therefore
It will be more accurate in conjunction with the estimation of figure, gender, ethnic group and empirical data.On the other hand, based on the artificial intelligence of deep learning
Identification technology rapid development makes it possible that we assess body shape using artificial intelligence and obtain related parameter.Therefore it adopts
The method for carrying out body fat assessment to user 1 with the above-mentioned body fat assessment system based on deep learning technology is as follows:
1) on intelligent Human fat balance 4, intelligent Human fat balance 4 measures the body fat of user 1, and will be measured at user 1 station
Data are transferred to client terminal machine 3, which is uploaded to cloud experience database 2 by client terminal machine 3, and with cloud experience number
It is compared according to original empirical data in library 2；
2) user 1 makes a series of dynamic actions, and client terminal machine 3 carries out the record of figure dynamic 3-D video to user 1, and
The figure dynamic 3-D video data are uploaded to cloud experience database 2, and with original experience in cloud experience database 2
Data compare；
3) data comparison that client terminal machine 3 is obtained according to step 1) and step 2 is as a result, obtain the body fat rate of user, muscle
Rate, protein content, bone amount and moisture content.
In the method, a series of dynamic actions that user 1 makes include turn left 90 degree, rear turnback, turn right 45 degree,
Bend over 90 degree, arm it is flattened.
Since convolutional neural networks (CNN) image recognition is the high image recognition technology of current accuracy, this hair
Client terminal machine in bright is when obtaining user's figure dynamic 3-D video, and takes the technology, specifically includes training process
With two stages of identification process, it is specific as follows:
Training process are as follows:
A. propagated forward process:
1, from taking a sample to be input in network in sample set；2, corresponding reality output is calculated.
In this stage, the information of input is transferred to output layer by successively transformation, mainly before to feature extraction.
B. back-propagation process:
1, the difference of reality output and desired output is calculated；2, by the method backpropagation of minimization error, weight matrix is adjusted；Instead
It is exactly the reverse feedback of error and the update of weight to propagation.
Training process specifically includes the following steps:
1. the sequence of user human body image initial data that pair client terminal machine obtains carries out pretreatment operation, including image reconstruction
And image enhancement；
2. the important area for influencing body fat assessment to a series of pretreated human body images extracts, individual profile diagram is obtained
Picture, skin surface image and gray scale, colouring information；
3. obtaining Human fat balance the data obtained includes the real bodies data such as height；
4. it is training data by a series of human body images and the biggish region group cooperation of impact evaluation numerical value weight, it is corresponding
The fat or thin actual conditions of user form training set as training label；
5. training network parameter in training data and label data input convolutional neural networks is first successively instructed using convolutional layer
Practice and extract feature, the method for reusing feedback is integrally finely tuned；
6. the trained network parameter of step 5 is saved, cloud experience database is uploaded, as human bioequivalence model.
Identification process are as follows:
1. input picture by filter and can biasing set carry out convolution obtain C1 layers；
2. couple C1 layers of characteristics of human body, which schemes progress down-sampling, obtains S2 layers；
3. couple S2 layers of characteristics of human body, which schemes progress convolution, obtains C3 layers；
4. couple C3 layers of characteristics of human body, which schemes progress down-sampling, obtains S4 layers；
The vector become after 5.S4 layers of characteristic pattern rasterisation is input to traditional full Connection Neural Network and is further identified,
It is exported；
Identification process specifically includes the following steps:
1. the sequence of user human body image initial data that pair client terminal machine obtains carries out pretreatment operation, including image reconstruction
And image enhancement；
2. obtaining Human fat balance the data obtained includes the real bodies data such as height；
3. by a series of human body image datas pre-processed and the input of intelligent Human fat balance the data obtained is trained is stored in visitor
The identification model of family terminating machine and/cloud experience database, obtains recognition result；
The convolution process of convolutional neural networks CNN is: 1), with a trainable filter deconvolute input picture or feature
Figure obtains convolutional layer then plus a biasing；2), down-sampling process: the several pixels in field become a picture by pond
Element, then by weighting, biasing is set, then by a Sigmoid function, generates Feature Mapping figure；Convolutional neural networks are basic
Structure includes two layers, and one is characterized extract layer, and the input of each neuron is connected with the local acceptance region of preceding layer, and extracts
The feature of the part, after the local feature is extracted, its positional relationship between other feature is also decided therewith；Its
Second is that Feature Mapping layer, each computation layer of network is made of multiple Feature Mappings, and each Feature Mapping is a plane, plane
The weight of upper all neurons is equal.Feature Mapping structure is using the small sigmoid function of influence function core as convolutional network
Activation primitive so that Feature Mapping have shift invariant.Further, since the neuron on a mapping face shares weight,
Thus reduce the number of network freedom parameter.Each of convolutional neural networks convolutional layer all followed by one is used to ask office
Portion is averagely and the computation layer of second extraction, this distinctive structure of feature extraction twice reduce feature resolution.
According to above embodiment, the present invention can be realized well.It is worth noting that based on said structure design
Under the premise of, to solve same technical problem, even if that makes in the present invention is some without substantive change or polishing, adopted
The essence of technical solution is still as the present invention, therefore it should also be as within the scope of the present invention.
Claims (7)
1. a kind of body fat assessment system based on deep learning technology, which is characterized in that whole including cloud experience database, client
Terminal and intelligent Human fat balance with skin temperature and humidity surveying function, the cloud experience database and the client terminal machine
Between establish data exchange channels, establish data exchange channels between the client terminal machine and the intelligent Human fat balance, it is described
Record has experience data in the experience database of cloud, and the empirical data is carried out according to the gender of people, age, figure and ethnic group
Classification, and including body fat rate, muscle rate, protein content, bone amount, moisture content and anatomical data；
The cloud experience database has the function of deep learning, can constantly acquire the new number different from original empirical data
According to, and by original empirical data in conjunction with new data after re-start classification.
2. the body fat assessment system based on deep learning technology according to claim 1, which is characterized in that the intelligence body fat
Scale is the Human fat balance based on BIA biologic resistance mensuration.
3. the body fat assessment system based on deep learning technology according to claim 2, which is characterized in that the intelligence body fat
The conductive film of scale is ITO conductive film, is embedded with thermocouple, can measure contacted skin temperature, which is used to adjust to adopt
The resistance value of sample.
4. the body fat assessment system based on deep learning technology according to claim 3, which is characterized in that the intelligence body fat
Moisture probe is installed inside scale, ambient humidity can be measured, which is used to adjust the resistance value of sampling.
5. the body fat assessment system based on deep learning technology according to claim 1, which is characterized in that the client terminal
Machine is smart phone or PC machine.
6. a kind of body fat appraisal procedure based on deep learning technology, which is characterized in that using 1,2,3,4 or 5 any one of power
The body fat assessment system based on deep learning technology carries out body fat assessment to user, comprising the following steps:
1) for subscriber station on intelligent Human fat balance, intelligent Human fat balance measures the body fat of user, and by measured data
Be transferred to client terminal machine, which is uploaded to cloud experience database by client terminal machine, and in the experience database of cloud
Original empirical data compare；
2) user makes a series of dynamic actions, and client terminal machine carries out the record of figure dynamic 3-D video to user, and will
The figure dynamic 3-D video data are uploaded to cloud experience database, and with original empirical data in the experience database of cloud
It compares；
3) client terminal machine is obtained according to step 1) and step 2 data comparison as a result, obtain the body fat rate of user, muscle rate,
Protein content, bone amount and moisture content.
7. the body fat appraisal procedure based on deep learning technology according to claim 6, which is characterized in that user make one
Serial dynamic action include turn left 90 degree, rear turnback, turn right 45 degree, bend over 90 degree, arm it is flattened.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910365560.XA CN110051353A (en) | 2019-05-01 | 2019-05-01 | A kind of body fat assessment system and method based on depth learning technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910365560.XA CN110051353A (en) | 2019-05-01 | 2019-05-01 | A kind of body fat assessment system and method based on depth learning technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110051353A true CN110051353A (en) | 2019-07-26 |
Family
ID=67322064
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910365560.XA Withdrawn CN110051353A (en) | 2019-05-01 | 2019-05-01 | A kind of body fat assessment system and method based on depth learning technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110051353A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113425254A (en) * | 2021-05-10 | 2021-09-24 | 复旦大学 | Youth male body fat rate prediction method based on body fat rate prediction model with mixed data input |
| US11232629B1 (en) | 2019-08-30 | 2022-01-25 | Amazon Technologies, Inc. | Two-dimensional image collection for three-dimensional body composition modeling |
| US11423630B1 (en) | 2019-06-27 | 2022-08-23 | Amazon Technologies, Inc. | Three-dimensional body composition from two-dimensional images |
| US11854146B1 (en) | 2021-06-25 | 2023-12-26 | Amazon Technologies, Inc. | Three-dimensional body composition from two-dimensional images of a portion of a body |
| US11861860B2 (en) | 2021-09-29 | 2024-01-02 | Amazon Technologies, Inc. | Body dimensions from two-dimensional body images |
| US11887252B1 (en) | 2021-08-25 | 2024-01-30 | Amazon Technologies, Inc. | Body model composition update from two-dimensional face images |
| US11903730B1 (en) * | 2019-09-25 | 2024-02-20 | Amazon Technologies, Inc. | Body fat measurements from a two-dimensional image |
-
2019
- 2019-05-01 CN CN201910365560.XA patent/CN110051353A/en not_active Withdrawn
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11423630B1 (en) | 2019-06-27 | 2022-08-23 | Amazon Technologies, Inc. | Three-dimensional body composition from two-dimensional images |
| US11232629B1 (en) | 2019-08-30 | 2022-01-25 | Amazon Technologies, Inc. | Two-dimensional image collection for three-dimensional body composition modeling |
| US11580693B1 (en) | 2019-08-30 | 2023-02-14 | Amazon Technologies, Inc. | Two-dimensional image collection for three-dimensional body composition modeling |
| US11903730B1 (en) * | 2019-09-25 | 2024-02-20 | Amazon Technologies, Inc. | Body fat measurements from a two-dimensional image |
| CN113425254A (en) * | 2021-05-10 | 2021-09-24 | 复旦大学 | Youth male body fat rate prediction method based on body fat rate prediction model with mixed data input |
| US11854146B1 (en) | 2021-06-25 | 2023-12-26 | Amazon Technologies, Inc. | Three-dimensional body composition from two-dimensional images of a portion of a body |
| US11887252B1 (en) | 2021-08-25 | 2024-01-30 | Amazon Technologies, Inc. | Body model composition update from two-dimensional face images |
| US11861860B2 (en) | 2021-09-29 | 2024-01-02 | Amazon Technologies, Inc. | Body dimensions from two-dimensional body images |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110051353A (en) | A kind of body fat assessment system and method based on depth learning technology | |
| CN105787439B (en) | A kind of depth image human synovial localization method based on convolutional neural networks | |
| CN108304826A (en) | Facial expression recognizing method based on convolutional neural networks | |
| CN108108807B (en) | Learning type image processing method, system and server | |
| CN109101938B (en) | Multi-label age estimation method based on convolutional neural network | |
| CN112307958A (en) | Micro-expression identification method based on spatiotemporal appearance movement attention network | |
| CN101819628B (en) | Method for performing face recognition by combining rarefaction of shape characteristic | |
| CN108021947B (en) | A kind of layering extreme learning machine target identification method of view-based access control model | |
| CN109902646A (en) | A kind of gait recognition method based on long memory network in short-term | |
| CN109308725A (en) | A kind of system that expression interest figure in mobile terminal generates | |
| CN105096279A (en) | Digital image processing method based on convolutional neural network | |
| Nattkemper et al. | A neural network architecture for automatic segmentation of fluorescence micrographs | |
| CN109389045A (en) | Micro- expression recognition method and device based on mixing space-time convolution model | |
| CN105809201A (en) | Identification method and device for autonomously extracting image meaning concepts in biologically-inspired mode | |
| CN110169754A (en) | Launch self-service skin detection system in public places | |
| CN107590478A (en) | A kind of age estimation method based on deep learning | |
| CN111178288B (en) | Human body posture recognition method and device based on local error layer-by-layer training | |
| CN106980830A (en) | One kind is based on depth convolutional network from affiliation recognition methods and device | |
| CN111090764A (en) | Image classification method and device based on multitask learning and graph convolution neural network | |
| CN108764123A (en) | Intelligent recognition human body sleep posture method based on neural network algorithm | |
| CN109567313B (en) | Intelligent insole with biological characteristic recognition function | |
| CN108492301A (en) | A kind of Scene Segmentation, terminal and storage medium | |
| CN114463675A (en) | Underwater fish group activity intensity identification method and device | |
| CN110363072A (en) | Tongue image recognition method, apparatus, computer equipment and computer readable storage medium | |
| CN112668486A (en) | Method, device and carrier for identifying facial expressions of pre-activated residual depth separable convolutional network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190726 |
|
| WW01 | Invention patent application withdrawn after publication |