CN109770913A - A kind of abnormal gait recognition methods based on reverse transmittance nerve network - Google Patents
A kind of abnormal gait recognition methods based on reverse transmittance nerve network Download PDFInfo
- Publication number
- CN109770913A CN109770913A CN201910064053.2A CN201910064053A CN109770913A CN 109770913 A CN109770913 A CN 109770913A CN 201910064053 A CN201910064053 A CN 201910064053A CN 109770913 A CN109770913 A CN 109770913A
- Authority
- CN
- China
- Prior art keywords
- gait
- abnormal
- recognition methods
- function
- data
- 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.)
- Pending
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention belongs to biometrics identification technology field, specially a kind of abnormal gait recognition methods based on reverse transmittance nerve network.Comprising: acquire human normal walking using the IMU for being worn on human body and simulate signal when typical abnormal gait is walked, the 3-axis acceleration information under different gaits is obtained;According to target representative row walk cadence by initial data do windowing cutting pre-process and each data queue is stamped by respective labels according to gait classification;Construct BNPP reverse transmittance nerve network;Obtained data label is sent into BPNN and be trained to training set and test set, training set is divided into, test set assessment models classifying quality is utilized after the completion of training.The present invention carries out Direct Classification to original I MU 3-axis acceleration data by increasing input layer node quantity, complicated gait cycle is eliminated to divide and feature extraction engineering, the classification accuracy to a variety of abnormal gaits is improved, data prediction workload is reduced, improves classification accuracy.
Description
Technical field
The invention belongs to biometrics identification technology fields, and in particular to abnormal gait recognition methods.
Background technique
Gait refers to the posture showed when people walks, is one of human body important biomolecule feature.Abnormal gait mostly with disease
Change position is related, as the important feature of reflection human health status and capacity, obtains accurate believable gait in time and believes
Breath, training abnormal gait classifier carries out timely early warning to abnormal gait, and it is monitored and is assessed for a long time, examines in medical treatment
There is important directive significance in disconnected, disease prevention.
The gait recognition method of mainstream mainly divides computer vision scheme based on video and image procossing and is based at present
The sensor plan of pavement and wearable sensor such as IMU.Two schemes all refer to a large amount of professional numbers after obtaining initial data
Data preprocess work and complicated Feature Engineering, to extract each correlated characteristic in gait cycle, although accuracy rate is high in real time
Property it is poor, computation complexity is high.And mainstream Gait Recognition system only provides each numerical index, the identification and classification task of abnormal gait
It is mainly completed by human expert, needs a large amount of professional knowledge of related fields.
Summary of the invention
The purpose of the present invention is to provide a kind of accuracy, and abnormal gait high, that real-time is good, computation complexity is low identifies
Method.
Abnormal gait recognition methods provided by the invention is based on reverse transmittance nerve network technology, to the letter of acquisition
Directly classified to abnormal gait by original signal after number being pre-processed, the specific steps are as follows:
Step 1 utilizes the IMU(inertia sensing unit for being worn on human body) acquisition human normal walking when signal, obtain normal
Gait 3-axis acceleration information;
Step 2 acquires signal when human body simulation typical case abnormal gait is walked using the IMU for being worn on human body, obtains abnormal step
State 3-axis acceleration information;
Step 3 walks cadence according to target representative row, initial data is done windowing cutting pretreatment, and will be each according to gait classification
Data queue stamps respective labels, forms data label to collection;
Step 4 is built BNPP reverse transmittance nerve network (BPNN), and BPNN network is three layers of full connection structure, and input layer is all
Node connectedness is to hidden layer, all Node connectedness of hidden layer to output layer;Define input layer, hidden layer, output layer and it is each swash
Function living;
Data label obtained in step 3 is sent into step 4 and is taken to training set and test set, training set is divided by step 5
The BPNN built is trained, and test set assessment models classifying quality is utilized after the completion of training.
The present invention Step 1: in step 2, the IMU acquisition unit can dispose but be not limited to shank, in shoes, loins, but
It is required that normal gait is consistent with the mounting means of abnormal gait.
In step 2 of the present invention, it is described typical case abnormal gait include but is not limited to hemiplegic gait, parkinsonian gait, duck step by step
State, steppage gait, diplegia gait, etc..
Described to do windowing cutting pretreatment to initial data in step 3 of the present invention, data window length representative value is two
A gait cycle, step value takes away a quarter of window length paragraph by paragraph between window, and the label form is but is not limited to one-hot coding.
In step 4 of the present invention, the input layer nodal point number NinputnodesEvery segment data window is defined as in step 3 current
Sample frequency fcUnder the sampling number from tri- axis of IMU, namely:
, wherein TwindowsFor data window time domain length.
The output layer nodal point number is defined as gait category number namely a kind of normal gait and all exceptions in "current" model
The sum of gait.
The weight optimization method of three layers of BPNN uses gradient descent method.
The hidden layer excitation function of the BPNN can be but be not limited to ReLU function, Sigmod function or tanh function, defeated
Layer excitation function is Softmax function out.
Further, the cost function form of the BPNN is that the form of L2 norm is added in calculating variance.
The invention has the advantages that constructing three layers of reverse transmittance nerve network, and by increasing input layer nodal point number
The mode of amount carries out Direct Classification to original I MU 3-axis acceleration data, and initial data only needs to carry out simply to open a window interception simultaneously
It is sent into neural network, so that complicated period division and feature extraction engineering are eliminated, at the same time to abnormal gait
The available raising of classification accuracy can reach 94% in more classification task preferred embodiments, reduce data prediction work
Amount improves classification accuracy and identification effect.
Detailed description of the invention
Fig. 1 is that the present invention is based on the abnormal gait recognition methods flow charts of reverse transmittance nerve network.
Fig. 2 is original signal example under normal gait in present pre-ferred embodiments (interception 10s).
Fig. 3 is original signal example (interception 10s) under hemiplegic gait in the embodiment of the present invention (drawing circle gait).
Fig. 4 is original signal example under parkinsonian gait (festinating gait) in the embodiment of the present invention (interception 10s).
Fig. 5 is original signal example (interception 10s) under gluteus medius myopathic gait (duck step by step state) in the embodiment of the present invention.
Fig. 6 is original signal example under steppage gait in the embodiment of the present invention (interception 10s).
Fig. 7 is original signal example under diplegia gait (scissor-like gait) in the embodiment of the present invention (interception 10s).
Fig. 8 is to be sent into signal example contained by a segment signal window of neural network every time in present pre-ferred embodiments.
Fig. 9 is the structure chart for the reverse transmittance nerve network that the present invention is built.
Figure 10 is confusion matrix of the model on test set after training in the embodiment of the present invention.
Figure 11 is reverse transmittance nerve network training flow chart of the invention.
Specific embodiment
Develop simultaneously preferred embodiment with reference to the accompanying drawing, and the present invention will be described in detail.
The present invention provides a kind of the human motion Approach for Gait Classification based on convolutional neural networks, method flow schematic diagram
As shown in Figure 1.This method is realized using following steps:
IMU hardware system is fixed on the outside of right leg, wherein Y-axis and horizontal plane, X-axis and people by step 1 by bandage
Body coronal-plane is vertical, and Z axis is vertical with human body sagittal plane.Setting systematic sampling rate is 512Hz, and setting IMU accelerometer is led sensitive
Degree is ± 2g, and motor message when acquisition human normal is walked, acquisition step number is no less than 100 steps, obtains three axis of normal gait and add
Velocity information, original signal example (interception 10s) is as shown in Figure 2 under one section of normal gait.
Step 2, IMU is placed to be accelerated with the same step 1 of collecting flowchart, tri- axis of IMU acquired under target simulation abnormal gait
Spend information.Choose five kinds of allusion quotations in abnormal gait because caused by from common causatives such as myopathy, osteopathy, brain damages in the present embodiment
Type abnormal gait is simulated, and is respectively hemiplegic gait (drawing circle gait), parkinsonian gait (festinating gait), gluteus medius myopathy
Gait (duck step by step state), steppage gait, diplegia gait (scissor-like gait).It is acquired under five kinds of abnormal gaits in the present embodiment
Original signal example (interception 10s) is as shown in Figure 3.
Step 3 walks cadence according to target representative row and initial data is done windowing cutting pretreatment, and according to gait classification
Each data queue is stamped into respective labels.Every segment signal window takes all sampled points in 2000ms in the present embodiment, at 512Hz
Available 3072 points of three axis.The present embodiment lifts six classification tasks, and number label is 0 to 5, and wherein hemiplegic gait (draws circle
Gait) label be 0, the label of parkinsonian gait (festinating gait) is 1, and the label of normal gait is 2, gluteus medius myopathic gait
The label of (duck step by step state) is 3, and the label of steppage gait is 4, and the label of diplegia gait (scissor-like gait) is 5.In two classification
In task (normal gait and abnormal gait), it is 1 that normal gait, which can re-flag, and it is 0 that all abnormal gaits, which can re-flag,.
Step 4 builds BNPP reverse transmittance nerve network.According to input data dimension (3*1024), input layer knot is set
Points are 3072.According to neural network model experience, it is 512 that hidden layer nodal point number, which is arranged, according to the present embodiment final classification number,
It is 6 that output layer nodal point number, which is arranged,.Hidden layer excitation function can be but be not limited to ReLU function, Sigmod function, tanh function, this
The linear amending unit of ReLU, formula are used in embodiment are as follows:
Output layer excitation function is Softmax function, formula are as follows:
Wherein, wjIt (j=0,1,2,3,4,5) is the weight vectors from hidden layer to output layer.The cost letter of convolutional neural networks
Number can be selected as conventionally form or cross entropy cost function form, and the cost function of convolutional neural networks is selected as intersecting in this implementation
Entropy cost function adds L2 regularization parameter, cross entropy concrete form are as follows:
Wherein,For i-th of value in label,For the respective components in the vector through softmax normalized output.L2
Canonical concrete form are as follows:
Wherein, EinIt is the training sample error for not including regularization term, λ is regularization parameter.BPNN network structure such as Fig. 9 institute
Show.
Step 5, by label obtained in step 3 switch to one-hot coding (10000,01000,001000,000100,
000010,000001), and by information window, label 65% is taken to do training set after corresponding, remaining 35% does test set.It will train
The BPNN that collection is sent into step 4 is trained and does ten folding cross validations.BPNN algorithm flow is as shown in figure 11.
In the present embodiment, batchsize namely every wheel train taken sequence sets number to take 128, and learning rate takes 0.001,
Regularization parameter takes 0.0001, and train epochs take 10000 steps to be trained.Classified after the completion of training using test set assessment models
Effect.
It is verified, in the present embodiment model test set six classification accuracy be 94.978%, F1 value (accurate rate and
The harmomic mean of recall rate) it is 96.092%.The confusion matrix of test result is as shown in Figure 10.
It should be pointed out that above-described embodiment is merely to illustrate the present invention, being achieved in that for each step can be
Variation, various modifications to these embodiments are it will be apparent that therefore all for those skilled in the art
The equivalents and improvement carried out on the basis of theory of the invention general and spirit, should all protection scope of the present invention it
It is interior.
Claims (9)
1. a kind of abnormal gait recognition methods based on reverse transmittance nerve network, which is characterized in that specific step is as follows:
Step 1 acquires signal when human normal is walked using the IMU for being worn on human body, obtains normal gait 3-axis acceleration
Information;
Step 2 acquires signal when human body simulation typical case abnormal gait is walked using the IMU for being worn on human body, obtains abnormal step
State 3-axis acceleration information;
Step 3 walks cadence according to target representative row, initial data is done windowing cutting pretreatment, and will be each according to gait classification
Data queue stamps respective labels, forms data label to collection;
Step 4 is built BNPP reverse transmittance nerve network (BPNN), and BPNN network is three layers of full connection structure, and input layer is all
Node connectedness is to hidden layer, all Node connectedness of hidden layer to output layer;Define input layer, hidden layer, output layer and it is each swash
Function living;
Data label obtained in step 3 is sent into step 4 and is taken to training set and test set, training set is divided by step 5
The BPNN built is trained;Test set assessment models classifying quality is utilized after the completion of training.
2. abnormal gait recognition methods according to claim 1, which is characterized in that Step 1: IMU described in step 2
Acquisition unit be placed in shank, in shoes, loins, normal gait is consistent with the mounting means of abnormal gait.
3. abnormal gait recognition methods according to claim 1, which is characterized in that typical case described in step 2 walks extremely
State includes hemiplegic gait, parkinsonian gait, duck state, steppage gait, diplegia gait step by step.
4. abnormal gait recognition methods according to claim 1,2 or 3, which is characterized in that original described in step 3
Beginning data do windowing cutting pretreatment, and data window length is two gait cycles, and piecewise step value takes away window length between window
A quarter, the label form are one-hot coding.
5. abnormal gait recognition methods according to claim 4, which is characterized in that input layer node described in step 4
Number is defined as sampling number of every segment data window from tri- axis of IMU under current sampling frequency in step 3;The output layer knot
Points are defined as gait category number in "current" model, i.e., the sum of a kind of normal gait and all abnormal gaits.
6. abnormal gait recognition methods according to claim 5, which is characterized in that three layers of BPNN's described in step 4
Weight optimization method uses gradient descent method;
The hidden layer excitation function of the BPNN can be ReLU function, Sigmod function or tanh function, output layer excitation function
For Softmax function.
7. abnormal gait recognition methods according to claim 6, which is characterized in that the cost letter of BPNN described in step 4
Number form formula is that the form of L2 norm is added in calculating variance.
8. abnormal gait recognition methods according to claim 6, which is characterized in that output layer excitation function is Softmax
Function, formula are as follows:
Wherein, wjIt (j=0,1,2,3,4,5) is the weight vectors from hidden layer to output layer.
9. abnormal gait recognition methods according to claim 7, which is characterized in that the cost function of convolutional neural networks selects
Add L2 regularization parameter, cross entropy concrete form for cross entropy cost function are as follows:
Wherein,For i-th of value in label,For the respective components in the vector through softmax normalized output;L2
Canonical concrete form are as follows:
Wherein, EinIt is the training sample error for not including regularization term, λ is regularization parameter.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910064053.2A CN109770913A (en) | 2019-01-23 | 2019-01-23 | A kind of abnormal gait recognition methods based on reverse transmittance nerve network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910064053.2A CN109770913A (en) | 2019-01-23 | 2019-01-23 | A kind of abnormal gait recognition methods based on reverse transmittance nerve network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109770913A true CN109770913A (en) | 2019-05-21 |
Family
ID=66501981
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910064053.2A Pending CN109770913A (en) | 2019-01-23 | 2019-01-23 | A kind of abnormal gait recognition methods based on reverse transmittance nerve network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109770913A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110299203A (en) * | 2019-08-12 | 2019-10-01 | 电子科技大学 | A kind of human body abnormal gait recognition methods based on smart phone |
CN111524537A (en) * | 2020-03-24 | 2020-08-11 | 苏州数言信息技术有限公司 | Method for identifying coughing and sneezing aiming at real-time voice flow |
CN111973193A (en) * | 2020-08-20 | 2020-11-24 | 中山大学 | Gait recognition method and system based on silicone-nickel nano sensor |
CN112818927A (en) * | 2021-02-26 | 2021-05-18 | 上海交通大学 | Real-time classification method and system for human body lower limb movement modes |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104091177A (en) * | 2014-06-30 | 2014-10-08 | 华南理工大学 | Abnormal gait detection method based on determined learning theory |
WO2016149881A1 (en) * | 2015-03-20 | 2016-09-29 | Intel Corporation | Object recogntion based on boosting binary convolutional neural network features |
CN108062170A (en) * | 2017-12-15 | 2018-05-22 | 南京师范大学 | Multi-class human posture recognition method based on convolutional neural networks and intelligent terminal |
CN108345846A (en) * | 2018-01-29 | 2018-07-31 | 华东师范大学 | A kind of Human bodys' response method and identifying system based on convolutional neural networks |
CN108958482A (en) * | 2018-06-28 | 2018-12-07 | 福州大学 | A kind of similitude action recognition device and method based on convolutional neural networks |
-
2019
- 2019-01-23 CN CN201910064053.2A patent/CN109770913A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104091177A (en) * | 2014-06-30 | 2014-10-08 | 华南理工大学 | Abnormal gait detection method based on determined learning theory |
WO2016149881A1 (en) * | 2015-03-20 | 2016-09-29 | Intel Corporation | Object recogntion based on boosting binary convolutional neural network features |
CN108062170A (en) * | 2017-12-15 | 2018-05-22 | 南京师范大学 | Multi-class human posture recognition method based on convolutional neural networks and intelligent terminal |
CN108345846A (en) * | 2018-01-29 | 2018-07-31 | 华东师范大学 | A kind of Human bodys' response method and identifying system based on convolutional neural networks |
CN108958482A (en) * | 2018-06-28 | 2018-12-07 | 福州大学 | A kind of similitude action recognition device and method based on convolutional neural networks |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110299203A (en) * | 2019-08-12 | 2019-10-01 | 电子科技大学 | A kind of human body abnormal gait recognition methods based on smart phone |
CN111524537A (en) * | 2020-03-24 | 2020-08-11 | 苏州数言信息技术有限公司 | Method for identifying coughing and sneezing aiming at real-time voice flow |
CN111524537B (en) * | 2020-03-24 | 2023-04-14 | 苏州数言信息技术有限公司 | Cough and sneeze identification method aiming at real-time voice flow |
CN111973193A (en) * | 2020-08-20 | 2020-11-24 | 中山大学 | Gait recognition method and system based on silicone-nickel nano sensor |
CN112818927A (en) * | 2021-02-26 | 2021-05-18 | 上海交通大学 | Real-time classification method and system for human body lower limb movement modes |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Cicirelli et al. | Human gait analysis in neurodegenerative diseases: a review | |
CN109770913A (en) | A kind of abnormal gait recognition methods based on reverse transmittance nerve network | |
CN109770912A (en) | A kind of abnormal gait classification method based on depth convolutional neural networks | |
CN109784412A (en) | The multiple sensor signals fusion method based on deep learning for gait classification | |
Kwolek et al. | Calibrated and synchronized multi-view video and motion capture dataset for evaluation of gait recognition | |
Yang et al. | Implementation of a wearerable real-time system for physical activity recognition based on Naive Bayes classifier | |
CN106580282A (en) | Human body health monitoring device, system and method | |
Li et al. | Abnormal sitting posture recognition based on multi-scale spatiotemporal features of skeleton graph | |
US20210020295A1 (en) | Physical function independence support device of physical function and method therefor | |
US20160128638A1 (en) | System and method for detecting and quantifying deviations from physiological signals normality | |
CN110659677A (en) | Human body falling detection method based on movable sensor combination equipment | |
Cleland et al. | Collection of a diverse, realistic and annotated dataset for wearable activity recognition | |
Dorschky et al. | Perspective on “in the wild” movement analysis using machine learning | |
Tao et al. | Attention-based sensor fusion for human activity recognition using imu signals | |
Kumar et al. | Human activity recognition (har) using deep learning: Review, methodologies, progress and future research directions | |
Zhen et al. | Hybrid Deep‐Learning Framework Based on Gaussian Fusion of Multiple Spatiotemporal Networks for Walking Gait Phase Recognition | |
Brzostowski et al. | Data fusion in ubiquitous sports training: methodology and application | |
CN108875501A (en) | Human body attribute recognition approach, device, system and storage medium | |
Low et al. | Lower extremity kinematics walking speed classification using long short-term memory neural frameworks | |
CN116543455A (en) | Method, equipment and medium for establishing parkinsonism gait damage assessment model and using same | |
CN111861275B (en) | Household work mode identification method and device | |
Zhou et al. | [Retracted] Analysis of Body Behavior Characteristics after Sports Training Based on Convolution Neural Network | |
Yashaswini et al. | Stress detection using deep learning and IoT | |
Wang et al. | Calorific expenditure estimation using deep convolutional network features | |
Kao et al. | Validation of automated mobility assessment using a single 3d sensor |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190521 |