CN109330602B - Female body intelligent evaluation detection device and method and storage medium - Google Patents
Female body intelligent evaluation detection device and method and storage medium Download PDFInfo
- Publication number
- CN109330602B CN109330602B CN201811295642.3A CN201811295642A CN109330602B CN 109330602 B CN109330602 B CN 109330602B CN 201811295642 A CN201811295642 A CN 201811295642A CN 109330602 B CN109330602 B CN 109330602B
- Authority
- CN
- China
- Prior art keywords
- human body
- detected
- sample
- point
- image
- 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.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
- A61B5/1077—Measuring of profiles
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
- A61B5/1071—Measuring physical dimensions, e.g. size of the entire body or parts thereof measuring angles, e.g. using goniometers
-
- 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
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Dentistry (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Fuzzy Systems (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a female body intelligent evaluation detection device and method, and a storage medium, wherein the female body intelligent evaluation detection method comprises the following steps: respectively acquiring the front side and the back side of a human body to be detected and a depth image to be detected in the walking process through a depth camera; acquiring a large number of depth images of the front side, the back side and the walking process of a human body, obtaining sample data of the depth images, establishing a female body shape evaluation model, obtaining a label image of the human body to be detected according to the female body shape evaluation model, and reconstructing morphological analysis calculation of the spine and the pelvis according to the label image of the human body to be detected to obtain a body shape evaluation report of the human body to be detected. The invention can perform non-invasive automatic evaluation on a plurality of key parts of the human body by using the depth camera and matching with the depth learning technology, thereby greatly shortening the inspection time and providing objective and quantized detection results for clinic.
Description
Technical Field
The invention relates to the field of human body shape assessment and detection, in particular to an intelligent female shape assessment and detection device and method and a storage medium.
Background
At present, the German DIERS product and the French KINEOD product use visible light to project the back of a human body, acquire two-dimensional pictures to perform three-dimensional simulation, analyze based on the characteristics of a single picture, and only analyze the shape of a spine. These two products have the following drawbacks:
1) the visible light projection has high requirements on the light of the use environment, and can be used in a light-tight environment, so that the input cost of a darkroom is increased;
2) although three-dimensional simulation reproduction is carried out through a two-dimensional picture, the three-dimensional simulation is only interface display and is not used for actual evaluation and judgment, because the two products are not provided with a depth camera for detection, and two-dimensional operation analysis is carried out through bright and dark pixels formed by visible light projection;
3) only a single picture is detected based on the pixel characteristics formed by the grating projection of visible light, the spine form is estimated, only the image processing and analysis are carried out, the deep learning technology is not used, and the accuracy of people with different forms is poor;
4) the two products only support spine morphology detection, and cannot detect other parts;
in the current clinical evaluation, a rehabilitation doctor or a doctor usually evaluates the form of a patient in a hand-test and visual-test mode, and the test time is long and the repeatability is poor. There are also clinicians who test patients by radioactive devices such as X-rays, which are costly, poorly instantaneous, and radiation-sensitive; in the existing postpartum rehabilitation examination, only part of doctors know the pelvic power model, the hands of postpartum patients can be visually measured, and structural abnormality of key parts of the body can be judged.
In particular, in postpartum women, according to the pelvic and abdominal dynamics theory, the appearance change caused by the birth may change the pressure state of the pelvic cavity and the abdominal cavity, and further clinical symptoms such as urinary incontinence and prolapse may be caused. Therefore, an early determination of the physical problem is necessary.
Disclosure of Invention
In view of the above technical problems, an object of the present invention is to provide an apparatus and a method for intelligently evaluating and detecting female body shape, and a storage medium, which solve the problems of low efficiency, high cost, and inaccurate detection result in body shape detection in the prior art.
The invention adopts the following technical scheme:
an intelligent evaluation and detection method for female body shapes comprises the following steps:
respectively acquiring the front side and the back side of a human body to be detected and a depth image to be detected in the walking process through a depth camera; the depth image in the walking process is obtained by taking one frame of image at intervals of time t for the dynamic video of human body walking;
acquiring a large number of depth images of the front side, the back side and the walking process of a human body to obtain sample data of the depth images, preprocessing the sample data to obtain sample point cloud data, and labeling the obtained sample point metadata to obtain sample label data; the labeling method is that a space coordinate value is set for all key points on the surface of a human body;
establishing a female body evaluation model, wherein the female body evaluation model takes sample data as input of a training sample and takes sample label data as output of the training sample;
inputting the acquired depth image to be detected into a female body evaluation model to obtain a tag image of the human body to be detected, wherein the tag image of the human body to be detected comprises space coordinate values of all key points on the surface of the human body;
reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating multiple body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected.
Further, the step of preprocessing the sample data to obtain sample point cloud data comprises:
calculating coordinates of points in the depth image corresponding to each pixel point by calling a Kinect API program to obtain a 3D point cloud image;
cutting the 3D point cloud image, wherein the coordinates of any point are (x, y, z), the height of a camera from the ground is h, and cutting off all points meeting y < -h to obtain a cut 3D point cloud image;
setting a coordinate threshold value of a limited space, taking a point of a central area of the cut 3D point cloud image, which is closest to the camera, as a seed point, and performing area growth to obtain a human body point element image segmented from the environment; and clearing the non-human body area in the human body point element image according to the mapping of the point cloud and the depth image to obtain sample point cloud data.
Further, the step of performing region growing specifically includes:
step A1, adding the seed points into a queue L;
step A2, judging whether there is some point in the queue L, if yes, taking out the point as P point, if not, executing step A3;
step A3, marking the P point, adding all unmarked points which are less than the threshold value d in the P point into a queue L, and continuing to execute the step A2;
step A3, keep all marked points and delete all unmarked points.
Further, the step of labeling the obtained sample point metadata to obtain sample label data specifically includes: and setting different label values for each point in the sample point cloud data according to the difference of the key points of the point on the human body surface, and setting the label value as a default value when the point does not belong to any key point on the human body surface.
Further, the step of labeling the obtained sample point metadata to obtain sample label data specifically includes:
marking the key points of the left and right acromions, the left and right anterior iliac spines, the pubic symphysis, the bilateral femoral greater trochanters, the medial upper ankle of the femur, the lateral upper ankle of the femur, the medial ankle of the lower limb and the lateral ankle of the lower limb of the obtained sample point metadata of the depth image of the front surface of the human body according to medical knowledge;
marking the key points of cervical vertebrae C1-C7, thoracic vertebrae T1-T12, lumbar vertebrae L1-L5, coccyx S5, lower marginal points of left and right shoulder blades, left and right posterior superior iliac spines, left and right fossa, bilateral femoral greater trochanter, internal upper ankle of femur, external upper ankle of femur, internal ankle of lower limb and external ankle of lower limb of the acquired depth image of the back of the human body and in the walking process according to medical knowledge.
Further, the step of establishing the female body evaluation model specifically comprises:
carrying out normalization preprocessing on the depth value of the depth image, using the processed depth value as the input of a model, and using sample label data as the output of a training sample;
performing iterative training on the model by using a training sample;
obtaining a key point coordinate with a label value T by calculating the mass center of all points with the label value T in the label graph;
continuously adjusting the network weight coefficient by using a random gradient descent method to continuously reduce the loss function until convergence; the loss function of the model is the root mean square of the Euclidean distance between the predicted key point and the actual key point; the accuracy of the model is the proportion of the prediction key point and the actual error mean value smaller than the threshold value d.
Further, the step of establishing the female body evaluation model further comprises:
acquiring a large number of verification samples of the depth images of the front side and the back side of the human body and in the walking process through a depth camera; and verifying the model by using the verification sample, wherein the model parameter with the highest accuracy of the model in the verification process is used as the model parameter of the model.
Further, reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating multiple body shape indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body shape evaluation report of the human body to be detected specifically comprises the following steps:
fitting each node (C1-S5) of the spine by using a high-order curve, and obtaining the tangent value of the included angle between a section of the spine containing the node and the gravity line by derivation of each point on the curve; the physiological curvature of the cervical vertebra is an included angle between a C1 node normal vector and a C5 node normal vector; the thoracic vertebra physiological curvature is an included angle between a normal vector of a T1 node and a normal vector of a T12 node; the physiological curvature of the lumbar vertebra is an included angle between a normal vector of an L1 node and a normal vector of an L5 node; the physiological curvature of the sacrum is an included angle between a normal vector of an S1 node and a horizontal plane; the length of the backbone of the vertebra is the sum of the connecting line distances of adjacent points C1-S5;
obtaining an included angle alpha1 between a connecting line of a left shoulder peak and a right shoulder peak and a horizontal line and an included angle alpha2 between a connecting line of a lower edge point of the left shoulder peak and the right shoulder blade and the horizontal line, if the alpha is max (i.e., | alpha1|, | alpha2|) is greater than a preset shoulder peak threshold value, judging that a high shoulder and a low shoulder exist, and judging which side is high and which side is low according to the positive and negative of alpha 1;
acquiring a detaY value of the difference of the longitudinal coordinates of the anterior superior iliac spines on both sides, and if the detaY value is greater than a preset threshold value of the iliac spines on both sides, judging that the pelvis is inclined leftwards or rightwards;
acquiring a plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and a human coronal plane, judging whether the pelvis rotates forwards or backwards according to whether the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis is parallel to the human coronal plane, and judging whether the pelvis rotates left and right according to the included angle and the direction of the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and the normal vector of the human coronal plane;
obtaining a femoral greater trochanter, an armpit, an inner ankle midpoint connecting line and an outer ankle midpoint connecting line to obtain a thigh trunk line and a shank trunk line, respectively comparing the angles of the left leg trunk line and the right leg trunk line which rotate anticlockwise relative to the thigh trunk line from the front to obtain an angle beta1 of the left leg trunk line relative to the thigh trunk line, and an angle beta2 of the right leg trunk line relative to the thigh trunk line; if the angle beta1 and the angle beta2 are both smaller than a preset leg threshold value, judging that the leg of the human body to be detected is normal; if the angle of beta1 is positive and the angle of beta2 is negative, the human body to be detected is judged to be a fork-shaped leg; if the angle of the beta1 is negative and the angle of the beta2 is positive, the human body to be detected is judged to be an O-shaped leg.
A female body shape evaluation device comprises a running machine, a depth camera and a processor, wherein the depth camera is used for acquiring depth images to be measured of a human body to be measured on the front side and the back side of the running machine in a walking process;
the processor is used for acquiring a large number of depth images of the front side, the back side and the walking process of a human body to obtain sample data of the depth images, preprocessing the sample data to obtain sample point cloud data, and labeling the obtained sample point metadata to obtain sample label data; the labeling method is that a space coordinate value is set for all key points on the surface of a human body; establishing a female body evaluation model, wherein the female body evaluation model takes sample data as input of a training sample and takes sample label data as output of the training sample; inputting the acquired depth image to be detected into a female body evaluation model to obtain a tag image of the human body to be detected, wherein the tag image of the human body to be detected comprises space coordinate values of all key points on the surface of the human body; reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating multiple body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected.
When the depth images to be detected of the front side, the back side and the walking process of a human body to be detected are obtained, the computer program is executed by a processor to realize the intelligent evaluation and detection method for the female body shape.
Compared with the prior art, the invention has the beneficial effects that:
by using the depth camera and matching with the depth learning technology, the invention can perform non-invasive automatic evaluation on a plurality of key parts of the human body, greatly reduce the inspection time and provide objective and quantized detection results for clinic; the traditional manual judgment method is intelligentized and automated through software, so that the postpartum physical abnormality assessment and screening can be conveniently popularized and popularized in the medical field.
Furthermore, the invention can carry out measurement and evaluation on various aspects such as pelvis, shoulders, lower limbs, overall shape and the like besides diagnosis on the shape and angle of the spine.
Drawings
FIG. 1 is a schematic flow chart of an intelligent evaluation and detection method for female body shape according to the present invention;
FIG. 2 is a schematic diagram of an intelligent evaluation and detection structure for female body shape according to the present invention;
wherein, 1, the treadmill standing line; 2. a camera center hole; 3. the front plane of the camera.
Detailed Description
The present invention will be further described with reference to the accompanying drawings and specific embodiments, and it should be noted that, in the premise of no conflict, the following described embodiments or technical features may be arbitrarily combined to form a new embodiment:
example (b):
referring to fig. 1, a method for intelligently evaluating and detecting female body shape includes the following steps:
s100, respectively acquiring the front side and the back side of a human body to be detected and depth images to be detected in the walking process through a depth camera; the depth image in the walking process is obtained by taking one frame of image at intervals of time t for the dynamic video of the human body walking;
s200, acquiring a large number of depth images of the front side, the back side and the walking process of a human body to obtain sample data of the depth images, preprocessing the sample data to obtain sample point cloud data, and labeling the obtained sample point metadata to obtain sample label data; the labeling method is that a space coordinate value is set for all key points on the surface of a human body;
further, the step of preprocessing the sample data to obtain sample point cloud data comprises: calculating coordinates of points in the depth image corresponding to each pixel point by calling a Kinect API program to obtain a 3D point cloud image;
cutting the 3D point cloud image, wherein the coordinates of any point are (x, y, z), the height of a camera from the ground is h, and cutting off all points meeting y < -h to obtain a cut 3D point cloud image;
setting a coordinate threshold value of a limited space, taking a point of a central area of the cut 3D point cloud image, which is closest to the camera, as a seed point, and performing area growth to obtain a human body point element image segmented from the environment; and clearing the non-human body area in the human body point element image according to the mapping of the point cloud and the depth image to obtain sample point cloud data.
Further, the step of performing region growing specifically includes:
step A1, adding the seed points into a queue L;
step A2, judging whether there is some point in the queue L, if yes, taking out the point as P point, if not, executing step A3;
step A3, marking the P point, adding all unmarked points which are less than the threshold value d in the P point into a queue L, and continuing to execute the step A2;
step A3, keep all marked points and delete all unmarked points.
Further, the step of labeling the obtained sample point metadata to obtain sample label data specifically includes: and setting different label values for each point in the sample point cloud data according to the difference of the key points of the point on the human body surface, and setting the label value as a default value when the point does not belong to any key point on the human body surface. Specifically, marking the key points of the left and right acromions, the left and right anterior iliac spines, the pubic symphysis, the bilateral femoral greater trochanters, the internal upper ankle of the femur, the external upper ankle of the femur, the internal ankle of the lower limb and the external ankle of the lower limb of the obtained sample point metadata of the depth image of the front surface of the human body according to medical knowledge;
marking the key points of cervical vertebrae C1-C7, thoracic vertebrae T1-T12, lumbar vertebrae L1-L5, coccyx S5, lower marginal points of left and right shoulder blades, left and right posterior superior iliac spines, left and right fossa, bilateral femoral greater trochanter, internal upper ankle of femur, external upper ankle of femur, internal ankle of lower limb and external ankle of lower limb of the acquired depth image of the back of the human body and in the walking process according to medical knowledge.
Step S300, establishing a female body shape evaluation model, wherein the female body shape evaluation model takes sample data as input of a training sample and takes sample label data as output of the training sample;
the method for establishing the female body evaluation model specifically comprises the following steps:
carrying out normalization preprocessing on the depth value of the depth image, using the processed depth value as the input of a model, and using sample label data as the output of a training sample;
performing iterative training on the model by using a training sample;
obtaining a key point coordinate with a label value T by calculating the mass center of all points with the label value T in the label graph;
continuously adjusting the network weight coefficient by using a random gradient descent method to continuously reduce the loss function until convergence; the loss function of the model is the root mean square of the Euclidean distance between the predicted key point and the actual key point; the accuracy of the model is the proportion of the prediction key point and the actual error mean value smaller than the threshold value d.
The step of establishing the female body evaluation model further comprises the following steps:
acquiring a large number of verification samples of the depth images of the front side and the back side of the human body and in the walking process through a depth camera; and verifying the model by using the verification sample, wherein the model parameter with the highest accuracy of the model in the verification process is used as the model parameter of the model.
Step S400, inputting the acquired depth image to be detected into a female body evaluation model to obtain a tag image of the human body to be detected, wherein the tag image of the human body to be detected comprises space coordinate values of all key points on the surface of the human body;
and S500, reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating multiple body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected.
Reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating a plurality of body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected specifically comprises the following steps:
fitting each node (C1-S5) of the spine by using a high-order curve, and obtaining the tangent value of the included angle between a section of the spine containing the node and the gravity line by derivation of each point on the curve; the physiological curvature of the cervical vertebra is an included angle between a C1 node normal vector and a C5 node normal vector; the thoracic vertebra physiological curvature is an included angle between a normal vector of a T1 node and a normal vector of a T12 node; the physiological curvature of the lumbar vertebra is an included angle between a normal vector of an L1 node and a normal vector of an L5 node; the physiological curvature of the sacrum is an included angle between a normal vector of an S1 node and a horizontal plane; the length of the backbone of the vertebra is the sum of the connecting line distances of adjacent points C1-S5;
obtaining an included angle alpha1 between a connecting line of a left and a right acromion and a horizontal line and an included angle alpha2 between a connecting line of a lower marginal point of the left and the right scapulae and the horizontal line, if the alpha is max (i.e., | alpha1|, | alpha2|) is larger than a preset acromion threshold value, judging whether a high shoulder or a low shoulder exists, and judging which side is high and which side is low according to the positive and negative of alpha 1;
obtaining a detaY value of the longitudinal coordinate difference of anterior superior iliac spines on both sides, and if the detaY value is larger than a preset threshold value of the iliac spines on both sides, judging that the pelvis is inclined leftwards or rightwards; if the left side is higher, the pelvis is right-leaning; if the right side is higher, the pelvis is inclined leftwards;
acquiring a plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and a human coronal plane, judging whether the pelvis rotates forwards or backwards according to whether the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis is parallel to the human coronal plane, and judging whether the pelvis rotates left and right according to the included angle and the direction of the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and the normal vector of the human coronal plane;
obtaining a femoral greater trochanter, an armpit, an inner ankle midpoint connecting line and an outer ankle midpoint connecting line to obtain a thigh trunk line and a shank trunk line, respectively comparing the angles of the left leg trunk line and the right leg trunk line which rotate anticlockwise relative to the thigh trunk line from the front to obtain an angle beta1 of the left leg trunk line relative to the thigh trunk line, and an angle beta2 of the right leg trunk line relative to the thigh trunk line; if the angle beta1 and the angle beta2 are both smaller than a preset leg threshold value, judging that the leg of the human body to be detected is normal; if the angle of beta1 is positive and the angle of beta2 is negative, the human body to be detected is judged to be a fork-shaped leg; if the angle of the beta1 is negative and the angle of the beta2 is positive, the human body to be detected is judged to be an O-shaped leg.
The female body shape evaluation device can be realized by adopting a device comprising a running machine, a depth camera and a processor, wherein the depth camera is used for acquiring depth images to be measured of a human body to be measured on the front side and the back side of the running machine and in the walking process; referring to FIG. 2, the treadmill station line 1 is parallel to the front plane 3 of the depth camera. The distance between the treadmill standing line and the front plane 3 of the depth camera is 200cm, the center hole of the depth camera is 95cm above the ground, and the height of the upper surface of the treadmill track is 3 cm. Of course, the female body shape evaluating apparatus of the present invention is not limited to the structure of the apparatus as long as the function of the detecting method of the present invention can be achieved.
The scoliosis symptom belongs to the category of bone surgery, has early clinical recognition, and has a very accurate detection method in equipment such as X-ray, MRI, CT and the like. Such non-invasive examinations by non-invasive external picture diagnosis are more efficient and less costly than conventional image examinations, and are suitable for clinical routine screening. The detection method can be used for male or female, and particularly can be used for detecting the body shape of the female after delivery caused by appearance change due to fertility and change of the pressure state of the pelvic cavity and the abdominal cavity.
In the invention, the depth information detection is realized by selecting the invisible infrared ray emitted by the depth camera, and the two-dimensional and three-dimensional models of the human body can be directly obtained by the depth camera without three-dimensional simulation by software. Further, my products usage data, in addition to still pictures, is dynamic video data. A core algorithm in software is used for acquiring more than 10 thousands of image data samples of a group by cooperation with numerous hospitals by applying a deep learning technology, performing characteristic analysis based on big data, obtaining an evaluation model by applying statistics and mathematics, training the model, adjusting and optimizing model parameters, and finally selecting a parameter product with high accuracy for use. Furthermore, besides diagnosis of the morphology and angle of the spine, measurement and evaluation of the morphology of the pelvis, shoulders, lower limbs and the whole body are also included. The examination time is greatly shortened, and objective and quantitative detection results are provided for clinics; the traditional manual judgment method is intelligentized and automated through software, so that the postpartum physical abnormality assessment and screening can be conveniently popularized and popularized in the medical field.
The present invention is a computer storage medium on which a computer program is stored, and the present invention can be stored in the computer-readable storage medium if it is implemented in the form of a software functional unit and sold or used as a stand-alone product. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may also be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.
Various other modifications and changes may be made by those skilled in the art based on the above-described technical solutions and concepts, and all such modifications and changes should fall within the scope of the claims of the present invention.
Claims (8)
1. An intelligent evaluation and detection method for female body shapes is characterized by comprising the following steps:
respectively acquiring the front side and the back side of a human body to be detected and a depth image to be detected in the walking process through a depth camera; the depth image in the walking process is obtained by taking one frame of image at intervals of time t for the dynamic video of the human body walking;
acquiring a large number of depth images of the front side, the back side and the walking process of a human body to obtain sample data of the depth images, preprocessing the sample data to obtain sample point cloud data, and labeling the obtained sample point cloud data to obtain sample label data; the labeling method is that a space coordinate value is set for all key points on the surface of a human body;
the step of labeling the obtained sample point cloud data to obtain sample label data specifically comprises the following steps:
marking key points of a left shoulder peak, a right anterior superior iliac spine, a pubic symphysis, bilateral femoral greater trochanters, an internal upper femoral ankle, an external upper femoral ankle, an internal lower limb ankle and an external lower limb ankle of the acquired sample point cloud data of the depth image of the front of the human body according to medical knowledge;
marking the key points of cervical vertebrae C1-C7, thoracic vertebrae T1-T12, lumbar vertebrae L1-L5, coccyx S5, lower marginal points of left and right shoulder blades, left and right posterior superior iliac spines, left and right fossa, bilateral femoral greater trochanter, internal upper ankle of femur, external upper ankle of femur, internal ankle of lower limb and external ankle of lower limb of the acquired depth image of the back of the human body and in the walking process according to medical knowledge;
establishing a female body evaluation model, wherein the female body evaluation model takes sample data as input of a training sample and takes sample label data as output of the training sample;
inputting the acquired depth image to be detected into a female body evaluation model to obtain a tag image of the human body to be detected, wherein the tag image of the human body to be detected comprises space coordinate values of all key points on the surface of the human body;
reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating multiple body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected;
reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating a plurality of body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected specifically comprises the following steps:
fitting each node (C1-S5) of the spine by using a high-order curve, and obtaining the tangent value of the included angle between a section of the spine containing the node and the gravity line by derivation of each point on the curve; the physiological curvature of the cervical vertebra is an included angle between a C1 node normal vector and a C5 node normal vector; the thoracic vertebra physiological curvature is an included angle between a normal vector of a T1 node and a normal vector of a T12 node; the physiological curvature of the lumbar vertebra is an included angle between a normal vector of an L1 node and a normal vector of an L5 node; the physiological curvature of the sacrum is an included angle between a normal vector of an S1 node and a horizontal plane; the length of the backbone of the vertebra is the sum of the connecting line distances of adjacent points C1-S5;
obtaining an included angle alpha1 between a connecting line of a left shoulder peak and a right shoulder peak and a horizontal line and an included angle alpha2 between a connecting line of a lower edge point of the left shoulder peak and the right shoulder blade and the horizontal line, if the alpha is max (i.e., | alpha1|, | alpha2|) is greater than a preset shoulder peak threshold value, judging that a high shoulder and a low shoulder exist, and judging which side is high and which side is low according to the positive and negative of alpha 1;
obtaining a detaY value of the longitudinal coordinate difference of anterior superior iliac spines on both sides, and if the detaY value is larger than a preset threshold value of the iliac spines on both sides, judging that the pelvis is inclined leftwards or rightwards;
obtaining a plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and a human coronal plane, judging whether the pelvis is inclined forwards or backwards according to whether the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis is parallel to the human coronal plane, and judging whether the pelvis rotates left and right according to the included angle and the direction of the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and the normal vector of the human coronal plane;
obtaining a femoral greater trochanter, an armpit, an inner ankle midpoint connecting line and an outer ankle midpoint connecting line to obtain a thigh trunk line and a shank trunk line, respectively comparing the angles of the left leg trunk line and the right leg trunk line which rotate anticlockwise relative to the thigh trunk line from the front to obtain an angle beta1 of the left leg trunk line relative to the thigh trunk line, and an angle beta2 of the right leg trunk line relative to the thigh trunk line; if the angle beta1 and the angle beta2 are both smaller than a preset leg threshold value, judging that the leg of the human body to be detected is normal; if the angle of beta1 is positive and the angle of beta2 is negative, the human body to be detected is judged to be a fork-shaped leg; if the angle of the beta1 is negative and the angle of the beta2 is positive, the human body to be detected is judged to be an O-shaped leg.
2. The method for intelligently evaluating and detecting female body shape according to claim 1, wherein the step of preprocessing sample data to obtain sample point cloud data comprises:
calculating coordinates of points in the depth image corresponding to each pixel point by calling a Kinect API program to obtain a 3D point cloud image;
cutting the 3D point cloud image, wherein the coordinates of any point are (x, y, z), the height of the camera from the ground is h, and cutting off all points meeting y < -h to obtain a cut 3D point cloud image;
setting a coordinate threshold value of a limited space, taking a point, closest to a camera, in a central area of the cut 3D point cloud image as a seed point, and performing area growth to obtain a human body point cloud image segmented from the environment; and clearing a non-human body area in the human body point cloud image according to the mapping of the point cloud and the depth image to obtain sample point cloud data.
3. The method for intelligently evaluating and detecting female body shapes according to claim 2, wherein the step of performing region growing specifically comprises:
step A1, adding the seed points into a queue L;
step A2, judging whether there is some point in the queue L, if yes, taking out the point as P point, if not, executing step A3;
step A3, marking the P point, adding all unmarked points which are less than the threshold value d in the P point into a queue L, and continuing to execute the step A2;
step A3, keeping all marked points and deleting all unmarked points.
4. The method for intelligently evaluating and detecting female body shapes according to claim 2, wherein the step of labeling the obtained sample point cloud data to obtain sample label data specifically comprises: and setting different label values for each point in the sample point cloud data according to the difference of the key points of the point on the human body surface, and setting the label value as a default value when the point does not belong to any key point on the human body surface.
5. The method for intelligently evaluating and detecting female body shapes according to claim 1, wherein the step of establishing the female body shape evaluation model specifically comprises the steps of:
carrying out normalization pretreatment on the depth value of the depth image, and taking the treated depth value as the input of a model and the sample label data as the output of a training sample;
performing iterative training on the model by using a training sample;
obtaining a key point coordinate with a label value T by calculating the mass center of all points with the label value T in the label graph;
continuously adjusting the network weight coefficient by using a random gradient descent method to continuously reduce the loss function until convergence; the loss function of the model is the root mean square of the Euclidean distance between the predicted key point and the actual key point; the accuracy of the model is the proportion of the prediction key point and the actual error mean value smaller than the threshold value d.
6. The method for intelligently evaluating and testing female body shapes according to claim 5, wherein the step of establishing the female body shape evaluation model further comprises the steps of:
acquiring a large number of verification samples of the depth images of the front side and the back side of the human body and in the walking process through a depth camera; and verifying the model by using the verification sample, wherein the model parameter with the highest accuracy of the model in the verification process is used as the model parameter of the model.
7. The female body shape evaluation device is characterized by comprising a running machine, a depth camera and a processor, wherein the depth camera is used for acquiring depth images to be measured of a human body to be measured on the front side and the back side of the running machine in a walking process;
the processor is used for acquiring a large number of depth images of the front side, the back side and the walking process of a human body to obtain sample data of the depth images, preprocessing the sample data to obtain sample point cloud data, and labeling the obtained sample point cloud data to obtain sample label data; the step of labeling the obtained sample point cloud data to obtain sample label data specifically comprises the following steps: marking key points of a left shoulder peak, a right anterior superior iliac spine, a pubic symphysis, bilateral femoral greater trochanters, an internal upper femoral ankle, an external upper femoral ankle, an internal lower limb ankle and an external lower limb ankle of the acquired sample point cloud data of the depth image of the front of the human body according to medical knowledge; marking the key points of cervical vertebrae C1-C7, thoracic vertebrae T1-T12, lumbar vertebrae L1-L5, coccyx S5, lower marginal points of left and right shoulder blades, left and right posterior superior iliac spines, left and right fossa, bilateral femoral greater trochanter, internal upper ankle of femur, external upper ankle of femur, internal ankle of lower limb and external ankle of lower limb of the acquired depth image of the back of the human body and in the walking process according to medical knowledge; the labeling method is that a space coordinate value is set for all key points on the surface of a human body; establishing a female body evaluation model, wherein the female body evaluation model takes sample data as input of a training sample and takes sample label data as output of the training sample; inputting the acquired depth image to be detected into a female body evaluation model to obtain a tag image of the human body to be detected, wherein the tag image of the human body to be detected comprises space coordinate values of all key points on the surface of the human body; reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating multiple body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected; reconstructing the shapes of the spine and the pelvis according to the label image of the human body to be detected, calculating a plurality of body indexes of the human body to be detected, and performing comparative analysis according to the obtained index values and a normal threshold value to obtain a body evaluation report of the human body to be detected specifically comprises the following steps:
fitting each node (C1-S5) of the spine by using a high-order curve, and obtaining the tangent value of the included angle between a section of the spine containing the node and the gravity line by derivation of each point on the curve; the physiological curvature of the cervical vertebra is an included angle between a C1 node normal vector and a C5 node normal vector; the thoracic vertebra physiological curvature is an included angle between a normal vector of a T1 node and a normal vector of a T12 node; the physiological curvature of the lumbar vertebra is an included angle between a normal vector of an L1 node and a normal vector of an L5 node; the physiological curvature of the sacrum is an included angle between a normal vector of an S1 node and a horizontal plane; the length of the backbone of the vertebra is the sum of the distances of the connecting lines of all the adjacent points C1-S5;
obtaining an included angle alpha1 between a connecting line of a left shoulder peak and a right shoulder peak and a horizontal line and an included angle alpha2 between a connecting line of a lower edge point of the left shoulder peak and the right shoulder blade and the horizontal line, if the alpha is max (i.e., | alpha1|, | alpha2|) is greater than a preset shoulder peak threshold value, judging that a high shoulder and a low shoulder exist, and judging which side is high and which side is low according to the positive and negative of alpha 1;
obtaining a detaY value of the longitudinal coordinate difference of anterior superior iliac spines on both sides, and if the detaY value is larger than a preset threshold value of the iliac spines on both sides, judging that the pelvis is inclined leftwards or rightwards;
acquiring a plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and a human coronal plane, judging whether the pelvis rotates forwards or backwards according to whether the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis is parallel to the human coronal plane, and judging whether the pelvis rotates left and right according to the included angle and the direction of the plane formed by the combination of the anterior superior iliac spines on the two sides and the pubis and the normal vector of the human coronal plane;
obtaining a femoral greater trochanter, an armpit, an inner ankle midpoint connecting line and an outer ankle midpoint connecting line to obtain a thigh trunk line and a shank trunk line, respectively comparing the angles of the left leg trunk line and the right leg trunk line which rotate anticlockwise relative to the thigh trunk line from the front to obtain an angle beta1 of the left leg trunk line relative to the thigh trunk line, and an angle beta2 of the right leg trunk line relative to the thigh trunk line; if the angle beta1 and the angle beta2 are both smaller than a preset leg type threshold value, judging that the leg type of the human body to be detected is normal; if the angle of beta1 is positive and the angle of beta2 is negative, the human body to be detected is judged to be a fork-shaped leg; if the angle of beta1 is negative and the angle of beta2 is positive, the human body to be detected is judged to be the O-shaped leg.
8. A computer storage medium on which a computer program is stored, wherein when acquiring depth images to be measured of the front, back and walking of a human body to be measured, the computer program is executed by a processor to implement the female body shape intelligent assessment detection method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811295642.3A CN109330602B (en) | 2018-11-01 | 2018-11-01 | Female body intelligent evaluation detection device and method and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811295642.3A CN109330602B (en) | 2018-11-01 | 2018-11-01 | Female body intelligent evaluation detection device and method and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109330602A CN109330602A (en) | 2019-02-15 |
| CN109330602B true CN109330602B (en) | 2022-06-24 |
Family
ID=65313583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811295642.3A Active CN109330602B (en) | 2018-11-01 | 2018-11-01 | Female body intelligent evaluation detection device and method and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109330602B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3745352B1 (en) * | 2019-05-31 | 2023-01-18 | presize GmbH | Methods and systems for determining body measurements and providing clothing size recommendations |
| CN110495889B (en) * | 2019-07-04 | 2022-05-27 | 平安科技（深圳）有限公司 | Posture evaluation method, electronic device, computer device, and storage medium |
| CN110353629B (en) * | 2019-07-16 | 2021-11-30 | 河南理工大学 | Intelligent evaluation system for electronic spine measurement |
| WO2021179230A1 (en) * | 2020-03-12 | 2021-09-16 | 南方科技大学 | Scoliosis detection model generating method and computer device |
| CN112767319B (en) * | 2020-12-30 | 2024-06-25 | 无锡祥生医疗科技股份有限公司 | Breast ultrasound imaging method, device and storage medium based on point cloud data segmentation |
| CN114224322B (en) * | 2021-10-25 | 2023-06-02 | 上海工程技术大学 | Scoliosis assessment method based on key points of human bones |
| CN115272313B (en) * | 2022-09-27 | 2022-12-30 | 广州辉博信息技术有限公司 | Muscle balance degree analysis method, system and equipment based on depth image |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2219836C1 (en) * | 2002-04-01 | 2003-12-27 | Сарнадский Владимир Николаевич | Method for determining shape and spatial orientation of human body trunk in walking |
| CN102314712A (en) * | 2011-09-23 | 2012-01-11 | 暨南大学 | Pelvis modeling and pelvis parameter measurement method based on three-dimensional imaging technology |
| CN106250867A (en) * | 2016-08-12 | 2016-12-21 | 南京华捷艾米软件科技有限公司 | A kind of skeleton based on depth data follows the tracks of the implementation method of system |
| CN107392086A (en) * | 2017-05-26 | 2017-11-24 | 深圳奥比中光科技有限公司 | Apparatus for evaluating, system and the storage device of human body attitude |
| CN107981863A (en) * | 2017-12-17 | 2018-05-04 | 石家庄学院 | Human body static state body image-pickup method and body condition evaluation method and device used |
| KR20180117272A (en) * | 2017-04-19 | 2018-10-29 | 최우진 | Method and apparatus for analyzing body shape and recognizing disorders using analysis of spatial coordinates, and computer program therefor |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8170306B2 (en) * | 2007-04-25 | 2012-05-01 | Siemens Aktiengesellschaft | Automatic partitioning and recognition of human body regions from an arbitrary scan coverage image |
| US9165199B2 (en) * | 2007-12-21 | 2015-10-20 | Honda Motor Co., Ltd. | Controlled human pose estimation from depth image streams |
-
2018
- 2018-11-01 CN CN201811295642.3A patent/CN109330602B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2219836C1 (en) * | 2002-04-01 | 2003-12-27 | Сарнадский Владимир Николаевич | Method for determining shape and spatial orientation of human body trunk in walking |
| CN102314712A (en) * | 2011-09-23 | 2012-01-11 | 暨南大学 | Pelvis modeling and pelvis parameter measurement method based on three-dimensional imaging technology |
| CN106250867A (en) * | 2016-08-12 | 2016-12-21 | 南京华捷艾米软件科技有限公司 | A kind of skeleton based on depth data follows the tracks of the implementation method of system |
| KR20180117272A (en) * | 2017-04-19 | 2018-10-29 | 최우진 | Method and apparatus for analyzing body shape and recognizing disorders using analysis of spatial coordinates, and computer program therefor |
| CN107392086A (en) * | 2017-05-26 | 2017-11-24 | 深圳奥比中光科技有限公司 | Apparatus for evaluating, system and the storage device of human body attitude |
| CN107981863A (en) * | 2017-12-17 | 2018-05-04 | 石家庄学院 | Human body static state body image-pickup method and body condition evaluation method and device used |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109330602A (en) | 2019-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109330602B (en) | Female body intelligent evaluation detection device and method and storage medium | |
| JP2021504061A (en) | 3D Medical Imaging Methods and Systems for Identifying Spine Fractures | |
| CN108309334B (en) | Data processing method of spine X-ray image | |
| CN106803251B (en) | The apparatus and method of aortic coaractation pressure difference are determined by CT images | |
| WO2020259600A1 (en) | A device, process and system for diagnosing and tracking of the development of the spinal alignment of a person | |
| US11915420B2 (en) | Method for obtaining an image biomarker that quantifies the quality of the trabecular structure of bones | |
| Michoński et al. | Automatic recognition of surface landmarks of anatomical structures of back and posture | |
| Balasooriya et al. | Intelligent brain hemorrhage diagnosis using artificial neural networks | |
| CN111784704B (en) | MRI hip joint inflammation segmentation and classification automatic quantitative classification sequential method | |
| CN109903277A (en) | A kind of scoliosis detection method based on polynomial curve fitting | |
| CN110010229A (en) | A kind of ultrasonic image analogy method based on generation confrontation network | |
| Potočnik et al. | Computerized detection and recognition of follicles in ovarian ultrasound images: a review | |
| CN114287915A (en) | Noninvasive scoliosis screening method and system based on back color image | |
| Lu et al. | The JNU-IFM dataset for segmenting pubic symphysis-fetal head | |
| Jang et al. | Leg-length discrepancy variability on standard anteroposterior pelvis radiographs: an analysis using deep learning measurements | |
| TWI399194B (en) | Semi-automatic knee cartilage mri image segmentation based on cellular automata | |
| Sergeant et al. | Same task, same observers, different values: the problem with visual assessment of breast density | |
| TW202001797A (en) | Internal edge detection system and method thereof for processing medical images | |
| Alves et al. | Sex-based approach to estimate human body fat percentage from 2D camera images with deep learning and machine learning | |
| CN115762787A (en) | Eyelid disease surgery curative effect evaluation method and system based on eyelid topological morphology analysis | |
| CN115546123A (en) | Prediction system for rupture risk of abdominal aortic aneurysm based on machine learning | |
| Sun et al. | A statistical model of spine shape and material for population-oriented biomechanical simulation | |
| Drukker et al. | Vp18. 07: First trimester scans: how much time does it take to acquire the crl and nt? | |
| Balasooriya et al. | Intelligent brain hemorrhage diagnosis system | |
| Tameem et al. | Morphological atlases of knee cartilage: shape indices to analyze cartilage degradation in osteoarthritic and non-osteoarthritic population |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |