SYSTEMS AND METHODS FOR A HAND HYGIENE COMPLIANCE CHECKING SYSTEM WITH EXPLAINABLE FEEDBACK

Abstract

Various embodiments of a system and associated method for evaluating handwashing protocol compliance are disclosed herein. The system considers temporal combinations of segments of a test handwashing routine to recognize components of a compliant handwashing routine within the test handwashing routine. The system determines whether the components of the handwashing routine have been adequately performed and provides actionable feedback to a user to improve adherence to handwashing routine guidelines.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This is a non-provisional application that claims benefit to U.S. Provisional Patent Application Ser. No. 63/041,746 filed 19 Jun. 2020, which is herein incorporated by reference in its entirety.

FIELD

The present disclosure generally relates to hygiene compliance system, and in particular to a system and associated method for monitoring handwashing hygiene practices and providing explainable feedback.

BACKGROUND

The Covid19 outbreak has affected the lives of billions of people in this world with over 2.79 million infections and 196,000 deaths as of Apr. 24, 2020. One of the primary modes of human to human transmission is through droplets from an infected person landing on the hands of a healthy individual, who in turn touches their face. An effective mode of defense against the SARS-Cov2 virus is to thoroughly wash hands regularly before touching one's face.

The World Health Organization (WHO) and Center for Disease Control (CDC) has published guidelines for proper handwashing, which includes six components (FIG. 1): i) take soap and rub in both the palms, ii) right palm over left dorsum with interlaced fingers, iii) left palm over right dorsum with interlaced fingers, iv) rotational rubbing of right thumb clasped in left palm, v) rotational rubbing of left thumb clasped in right palm, vi) rotational rubbing backwards and forwards while rinsing. The entire duration of hand washing exercise should last for at least 20 seconds. Adherence to this protocol is necessary to eliminate the virus from ones' hands. However, adhering to these guidelines can be a challenge for many, especially when done frequently throughout the day.

It is with these observations in mind, among others, that various aspects of the present disclosure were conceived and developed.

BRIEF DESCRIPTION OF THE DRAWINGS

The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

FIG. 1 is a series of photographs showing WHO/CDC recommended components of a compliant handwashing routine;

FIG. 2 is a simplified diagram showing an embodiment of a hand hygiene monitoring system;

FIG. 3 is a simplified block diagram showing an additional view of the system of FIG. 2;

FIG. 4 is a diagram showing generation of models for each handwashing routine component of the system of FIG. 2;

FIG. 5 is a diagram showing identification of handwashing routine components and generation of explanations for missed or improperly completed handwashing routine components of the system of FIG. 2;

FIG. 6 is a process flow diagram showing a process for hand hygiene monitoring by the system of FIG. 2;

FIG. 7 is a diagram showing criteria for compliant hand washing data;

FIG. 8 is a diagram showing criteria for non-compliant hand washing data;

FIGS. 9A and 9B are graphical representations showing examples of RuLSIF results for compliant hand washing routines; and

FIG. 10 is a simplified diagram showing an exemplary computing system for implementation of the system of FIG. 2.

Corresponding reference characters indicate corresponding elements among the view of the drawings. The headings used in the figures do not limit the scope of the claims.

DETAILED DESCRIPTION

Monitoring hand hygiene practices, feedback on hygiene evaluation, and immediate action are important steps toward protection against a virus that transmits through droplets such as SARS-Cov2. A proper handwashing routine, one of the categories of hand hygiene, includes several components that must be performed in a temporally aligned sequence as recommended by WHO/CDC. The present system is an armband-based hand wash compliance checking mechanism that can identify if a component of the WHO/CDC recommendations are missing during a hand washing routine. An appropriate alert mechanism can then inform the user to redo the missing component(s) of the handwashing routine based on a compliance factor provided by the present system. Tests on data collected from 20 subjects, show a ten-fold user-independent accuracy of detecting compliant hand washing at 88%, precision at 0.89 and recall at 0.85. The present system has been found to identify any missing components in about 350 out of 500 non-compliant hand wash routines. Referring to the drawings, embodiments of a system and associated process for hand hygiene compliance checking are illustrated and generally indicated as 100 and 200 in FIGS. 1-10.

In this disclosure, an armband-based hand hygiene compliance checking system 100 (hereinafter “system”) is disclosed that evaluates user compliance with WHO/CDC handwashing guidelines. The system 100 further includes an explanation framework to alert a user of a potential reason for non-compliance, so that the user can immediately correct themselves.

Challenges

Given that the present WHO/CDC compliant hand washing routine includes six well-defined components, there can still be significant variations in correct handwashing routines performed by individuals. First, there is no specific order in which the components must be executed. The first step, taking soap, and the last step, rinsing, should be fixed, however, the rest of the components can occur in any order. Moreover, there is no restriction on the number of times each component is performed. Hence, a user may perform the same component many times during the hand washing routine, resulting in potentially infinite ways of following the WHO/CDC guidelines. As such a classification approach that considers compliant and non-compliant hand washing examples and trains a machine to verify compliance can be cumbersome and also infeasible.

Secondly, in case a user misses a given component of the hand washing routine, the user should be immediately notified so that the user can correct themselves. This necessitates an explainable compliance checking system that can identify the reason a particular handwashing routine is non-compliant.

Thirdly, the system is intended to scale to many users within a facility. As such it is not feasible to collect training data from each user. Hence, the compliance checking system 100 should be user independent.

Approach

The present system 100 is an armband-based hand hygiene monitoring system. In one aspect, the user wears an armband 104 which connects to a smartphone or other computing device 102 via Bluetooth or another suitable wireless or wired connection. The armband 104 collects motion data from onboard sensors 112 when a user is executing the handwashing routine. The user does not need to continuously wear the armband 104, it is required to be worn when performing a hand washing routine.

To address the challenges mentioned above, the WHO/CDC recommendations for a compliant hand washing routine can be interpreted as a language, which can generate an infinite number of compliant hand washing sequences of routine components. Each component of the compliant handwashing routine may be considered a letter in the alphabet of the language described by the WHO/CDC recommendations. The alphabet is the set of all steps in the recommendation. A given hand washing routine then needs to match a language expressed using a context free grammar.

For the present system, individual component models must first be developed for each component of the compliant hand washing routine. Data 171 from a compliant hand washing routine using armband 104 is recorded and then segmented into a plurality of segments of the routine using a change point detection module. Each segment of the plurality of segments is then checked against models of all the components of the compliant hand washing routine to determine if all components are present in an order enforced by a grammar (Eq. 1, below). This ensures that the user can perform all the components any number of times in any sequence. In the present embodiment, the system 100 assumes that the first and the sixth components of the routine are fixed, however, in other embodiments, alternative temporal arrangement of components may be required.

This component-wise recognition and language-based expression allows the present system 100 to identify a potential reason for noncompliance and provide feedback to the user.

To address the user-independence requirements, the present system 100 considers differential features of a given component of a handwashing routine. The differential feature captures the relative differences in the alignment of the arms with respect to the first and the sixth step of the routine.

Results Overview

Compliant and non-compliant hand washing examples were collected from 20 users for a total of 1300 samples. The system 100 reached a 10-fold cross-validation accuracy of 88% with a precision of 0.89 and recall of 0.85. This is with 60 (12 users)-40 (8 users) train test split among users. For non-compliant handwashing routines, the system 100 was able to identify the missing component in 75% of the instances. Once a non-compliant hand washing routine is detected by the system 100, the system 100 then asks the user to redo only the component(s) that were identified as missing.

Problem Statement

Each component of the WHO/CDC recommended handwashing routine can be modeled as a respective letter S_i, i∈1 . . . 6 in an alphabet Σ. The WHO/CDC recommendation can be encoded in the following context-free grammar:

S_2R→S₃*S₄*S₅*S₂S₂*S₃*S₄*S₅*

S_3R→S₂*S₄*S₅*S₃S₃*S₂*S₄*S₅*

S_4R→S₂*S₃*S₅*S₄S₄*S₂*S₃*S₅*

S_5R→S₂*S₃*S₄*S₄S₄*S₂*S₃*S₄*

H and Wash→S₁S_2RS_3RS_4RS_5RS₆  (1)

Here S_i∀i∈{1, 2, 3, 4, 5, 6} are symbols representing each component of the recommended routine. An asterisk (*) as a superscript on a symbol denotes a Kleene start operation, which means any number of occurrences of the symbol can occur within the handwashing routine including no occurrence. HandWash.dur denotes the duration of the hand washing routine. According to the current WHO/CDC recommendations at the time of this filing, HandWash.dur≥20 s.

The system 100 represents each handwashing routine in the form of a context free grammar, such as the context-free grammar described in Equation 1. The context-free grammar allows concepts or components of the routine to be represented in a modular fashion; that is, the definition allows flexibility in repetition or order of some components of the handwashing routine. Further, the context-free grammar enables compliance checking of handwashing routines even if the components of the routine are repeated, out of order, or have an extended duration.

Problem Statement: Given sets of training samples TS_ii∈{1 . . . 6}, for each handwashing component Si,

System Input: Test data from a handwashing routine including accelerometer data and inertial measurement unit (IMU) data.

System Outputs: 1. Identify whether the test data matches grammar of Equation 1, 2 that denote an acceptable handwashing routine. If the test data does not match the acceptable handwashing routine then determine which component of the grammar is missing. and 3. Duration of handwash.

System Model

An embodiment of the system 100 is shown in FIGS. 2 and 3. In one embodiment, the armband 104 is a Myo device, developed by Thalmic Labs. The armband 104 is a band-shaped gesture control device that is worn on an arm of a user. The armband 104 determines the motion of the arm and transmits a live feed via Bluetooth or another wireless or wired connection protocol to a connected device 102 such as a smartphone in order to access raw EMG and/or IMU data streamed at 200 Hz and 50 Hz respectively. In one embodiment, the armband 104 includes a plurality of sensors 112 that include eight surface ElectroMyoGraphy (EMG) sensors and a highly sensitive 9-axis Inertial Measurement Unit (IMU) sensor 114 that includes a 3-axis gyroscope, a 3-axis accelerometer, and a 3-axis magnetometer. In some embodiments, the EMG sensor is not necessary for operation of the system 100. This is because EMG sensors are currently not available in commercial wristbands or smart watches. Hence, if one wants to deploy the system 100 using a commercial wristband or smartwatch, the system 100 cannot rely exclusively on EMG. The armband 104 can include network interface 116 that communicates data to the computing device 102 across the network. Referring to FIG. 3, the computing device 102 includes a processor 150 in communication with a memory 120 and a network interface 160. The network interface 160 receives sensor data from the armband 104 for processing by the computing device 102. The memory 120 includes instructions for an operating system 140 and instructions for execution of a compliance checking processes/services application 130 corresponding with process 200.

In some embodiments, data is recorded from the armband 104 through an associated data recordation application. In some embodiments, the data recordation application is Myo Recode. The app collects the data from the sensors 112 and stores them in the app folder as five individual text (.txt) files i.e., EMG.txt, IMU.txt, location.txt, process.txt and service.txt as shown in FIG. 2.

Compliance Detection

Referring to FIGS. 3-6, the computing system 102 of the system 100 includes three modules: a) a compliance definition module 170 including models for each component of the WHO/CDC recommended hand washing routine, b) a component identification module 180 operable to identify components in a handwashing routine performed by a user based on the models in the compliance definition module 170, and c) a compliance checking and feedback module 190.

Modeling Each Component

Referring to FIG. 4, training data examples 171 of compliant handwashing routines were used to develop the compliance definition module 170. The start and end time stamps of each component of the compliant handwashing routing examples are labeled. Training data including orientation X, Y and accelerometer X, Y data from the armband 104 were considered for the examples. Differential features are extracted from each compliant handwashing example as shown in FIG. 4. For a given component S_i, a relative alignment of S_i is considered with respect to all other components. The alignment was computed using a Dynamic Time Warping (DTW) distance measure by a DTW module 183. Hence, for each gesture component S_i there are five DTW-based alignments from each mode of sensing. Hence in total there are 20 such alignments, which forms a feature vector for a given handwash component. Such feature vectors are extracted for all components S₂, S₃, S₄, and S₅.

In one implementation, to build a model of one component S_i of the compliant routine, a Support Vector Machine (SVM) 184 is trained with alignment feature vectors for S_i as a positive class and alignment feature vectors from S_i, j≠i as a negative class. Each routine component S_i is associated with an individual SVM 184 The Radial Basis Function was used as the kernel. To maintain balanced data, if one had x_i number of positive samples for S_i, the system 100 considers x_i number of negative class samples using a uniform distribution. The trained SVM 184 encodes a model for a handwash routine component S_i. In some embodiments, there are a plurality of differential SVM models 184A-D for components S₂, S₃, S₄, S₅. In some embodiments, the system 100 does not require SVM models of S₁ and S₆ as they are already encoded in the other models, given that they are used as references, however other embodiments may have other constraints on handwashing routine component arrangements.

Such differential models enable the system 100 to perform user-independent recognition. The primary reason is that the alignment operation using DTW ensures that artifacts related to initialization and origin shift are negated by the DTW operation.

Identifying Components

For deployment of the system 100, the component identification module 180 is shown in FIG. 5 and a corresponding process flow 200 is illustrated in FIG. 6. Referring to block 210 of process 200, given test data 181 of a handwashing routine captured by sensors 112, the system 100 first identifies segments including a number of segments within the test data 181 using a change point detection module 182. In particular, the change point detection module 182 can include a change point detection algorithm such as RuL-SIF, which is a divergence-based change point detection algorithm. The change point detection module 182 first divides the entire sample into windows. A window size is selected; one example window size is 4 seconds, i.e. 200 data points given a 50 Hz sampling ratio of the armband. The window size can be selected empirically; it was found that on average each component will take around 4 seconds to complete for a 20 second handwashing routine. The change point detection module 182 then computes a Pearson's divergence coefficient between two consecutive windows. The windows are then slid across the entire sample, hence obtaining a variation of the Pearson's Divergence coefficient over time. Peaks in a divergence coefficient indicates that there is a change in the underlying process generating the data. The change point detection module 182 extracts peaks from the Pearson's Divergence coefficient and the peak locations are considered to be start and end points of segments of test data 181. Based on these change points, the component identification module 180 segments the handwash test sample.

For this particular application, the system 100 assumes that a first segment of the test data 181 corresponds to S₁, i.e. take soap, and the last segment corresponds to S₆, i.e. rinse. Referring to block 220, the system 100 then computes relative DTW alignments of consecutive segments at a DTW module 183. The component identification module 180 considers all possible sequences of each segment of the test data 181 by temporally swapping one segment with another segment. For each temporal combination of segments, the component identification module 180 computes DTW alignments of consecutive segments. Referring to block 230, each handwash sample generates a feature vector of length at-least 120, (4C2×5) assuming that each segment is performed once.

Compliance Checking and Generating Explanation

The compliance checking and feedback module 190 evaluates the segments of the test data 181 to ensure that the hand washing routine is in compliance with the recommended hand washing routine. At block 240, the feature vector for each segment is passed to each of the four SVM component models 184A-D to classify segments of the test data 181 into components of a compliant handwashing routine. In particular, each temporal combination of segments is compared against each SVM component model 184 of a compliant routine. Each temporal combination of segments is passed through the SVM models 184. Each temporal combination results in a 4-tuple of SVM outputs. Hence, for each test handwash routine there will be at least 24 instances of SVM model outputs. At the end of this step, the system 100 obtains at least a 24×4 matrix. At block 250, the compliance checking and feedback module 190 obtains aggregate results from the SVM component models 184 to evaluate the SVM outputs This matrix is then summed up by the rows to get aggregate results of all combinations to yield a 1×4 matrix. These are the aggregate outputs of the SVM component models 184A-D. The i^th element of the matrix signifies the detection of the i^th component.

At block 250, the aggregate outputs of the SVM component models 184A-1840 are evaluated to identify a percentage of positive detections across the plurality of temporal combinations of segments. The percentage of positive detections for each SVM component model 184 is compared with a threshold percentage to determine whether the corresponding component of the compliant routine is sufficiently present within the test data. A threshold of 3% was set, which means that the component S_i should be recognized by the SVM model 184 for S_i in at least 30% of the temporal combinations. When all SVMs 184 return a positive class, then the compliance checking and feedback module 190 considers that the handwashing sample routine is compliant. If at least one SVM 184 returns a negative result, the compliance checking and feedback module 190 considers that the hand washing routine is non-compliant. At block 260, the system 100 compares an aggregation recognition result for test data 181 with a threshold to identify missing routine components that do not meet the threshold. If the aggregate recognition results are lower than this threshold for a component S_i, then the compliance checking and feedback module 190 considers that that component S_i was missing and at block 270 asks the user to redo that component.

Evaluation

Evaluation of the system 100 has two aims: a) determine the accuracy of identification of compliant hand washing routines, and b) determine the correctness of the explanation generation mechanism.

Evaluation Metrics

To address evaluation goals there are two sets of metrics. For the first aim, accuracy, precision, and recall are considered. In the specific application of the system 100 discussed herein, it is believed that precision is a more important parameter than recall for public safety. It is necessary to reduce false positives i.e. non-compliant hand washing routines classified as compliant. If compliant handwashing routines are classified as noncompliant, then there is no harm in spending an additional few seconds re-doing the routine.

For the second aim, the evaluation considers among how many instances the reason for non-compliance was correctly identified over the total number of non-compliant instances.

Data Collection

Data was collected from 20 subjects including correct and incorrect hand washings. Compliant handwashing is defined as the hand washes that have all the 6 required step actions performed for approximately 20 seconds either (a) sequentially or (b) in different combination orders with fixed step 1 and step 6 being the start and finish steps. Additionally, the data included correct handwashings lasting for approximately 40 seconds, twice the recommended amount of time.

As shown in FIG. 7, in criteria (a), each step action is performed for 4 times. In criteria (b), each step action is performed for 4 times having different sequences of steps keeping the first step 1 and last step 6 at the start and finish, hence there are 24 different combinations of handwashing styles i.e., S_i, different step orders of S₂, S₃, S₄, S₅, S₆. In criteria (c), every step is performed 7 times.

Non-compliant hand washing is defined as the hand washes that have performed for less than 15 seconds having (i) all the required steps sequentially or (ii) either one component/step is missing with different combination order of steps and (iii) handwashing that have one missing step action with different combination order of steps performed for 20 seconds. Missing component is between the steps 2 to 5.

As shown in FIG. 8, in criteria (i), every step action is performed 2 times. In criteria (ii), every step action is performed 2 times with one missing component of step and different order sequence of steps keeping the first step 1 and last step 6 at the start and finish i.e., S₁, missing of S₂ or S₃ or S₄ or S₅ and different step orders non-missing steps, S₆. In criteria (iii), it is similar fashion as criteria (ii) but every action is performed 5 or 6 times to get the 20 seconds duration.

Results

Segmentation is an important component of the entire mechanism. RuLSIF change point detection algorithm was used for segmentation and its accuracy was verified on the collected data. RuLSIF was 93% accurate in determining the change point, which means that out of nearly 8000 change points, it identified nearly 7500 change point time stamps. Out of the 7,500 time stamps, on an average RuLSIF based segmentation was off by +/−13 samples. This error may have contributed to lower the accuracy of our mechanism. FIGS. 9A and 9B show an example output of RuISIF based change point detection algorithm on the Orientation X axis of a compliant handwash routine.

Ten-fold cross-validation inter-user recognition was performed while considering a 60-40 split of train vs. test data. This implies that 12 users were used to train the system and the rest 8 users were used to test. The average accuracy for recognition of compliant hand washing among the total number of instances was 88%. The mechanism achieved an average precision of 0.89, while the recall was 0.85. A higher precision is an encouraging result since it indicates that false positives are lesser than false negatives.

Intra-user recognition with a 60-40 split was performed. This means that out of the 65 samples collected from each user, the evaluation included 60% data of each user in training and 40% from each user for testing. The accuracy increased to 92% and the precision was 96% while the recall was 90%. This is expected since performed user dependent analysis was also performed. However, the difference in accuracy is only 7% which indicates that the system 100 was successful in extracting user-independent characteristics of handwash routines.

For detection of missing components, the evaluation only focused on the non-compliant handwash routines. Duration of hand wash routine is easy to identify by just checking the number of data points in each routine. Hence, if a non-compliant handwash violates the duration rule, the system 100 can identify it 100% of the time. For the handwash routines that are more than 20 seconds, the system 100 attempts to identify the missing components. The dataset includes nearly 500 samples of non-compliant handwashing routines with one or more missing components. It is assumed that the explanation for non-compliance is successful if the system 100 can identify at least one component that is missing. Out of the 500 non-compliant samples the system 100 could correctly identify at least one missing component in 75% of the samples. The system 100 could correctly identify two missing components in 41% of the samples.

Computer-Implemented System

FIG. 10 illustrates an example of a suitable computing and networking environment (computer system 300) which may be used to implement various aspects of the present disclosure. Example embodiments described herein may be implemented at least in part in electronic circuitry; in computer hardware executing firmware and/or software instructions; and/or in combinations thereof. Example embodiments also may be implemented using a computer program product (e.g., a computer program tangibly or non-transitorily embodied in a machine-readable medium and including instructions for execution by, or to control the operation of, a data processing apparatus, such as, for example, one or more programmable processors or computers). A computer program may be written in any form of programming language, including compiled or interpreted languages, and may be deployed in any form, including as a stand-alone program or as a subroutine or other unit suitable for use in a computing environment. Also, a computer program can be deployed to be executed on one computer, or to be executed on multiple computers at one site or distributed across multiple sites and interconnected by a communication network.

Certain embodiments are described herein as including one or more modules. Such modules are hardware-implemented, and thus include at least one tangible unit capable of performing certain operations and may be configured or arranged in a certain manner. For example, a hardware-implemented module may comprise dedicated circuitry that is permanently configured (e.g., as a special-purpose processor, such as a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC)) to perform certain operations. A hardware-implemented module may also comprise programmable circuitry (e.g., as encompassed within a general-purpose processor or other programmable processor) that is temporarily configured by software or firmware to perform certain operations. In some example embodiments, one or more computer systems (e.g., a standalone system, a client and/or server computer system, or a peer-to-peer computer system) or one or more processors may be configured by software (e.g., an application or application portion) as a hardware-implemented module that operates to perform certain operations as described herein.

Accordingly, the term “hardware-implemented module” encompasses a tangible entity, be that an entity that is physically constructed, permanently configured (e.g., hardwired), or temporarily configured (e.g., programmed) to operate in a certain manner and/or to perform certain operations described herein. Considering embodiments in which hardware-implemented modules are temporarily configured (e.g., programmed), each of the hardware-implemented modules need not be configured or instantiated at any one instance in time. For example, where the hardware-implemented modules comprise a general-purpose processor configured using software, the general-purpose processor may be configured as respective different hardware-implemented modules at different times. Software, in the form of the system 100 or process 200 or otherwise, may include a hardware-implemented module and may accordingly configure a processor 302, for example, to constitute a particular hardware-implemented module at one instance of time and to constitute a different hardware-implemented module at a different instance of time.

Hardware-implemented modules may provide information to, and/or receive information from, other hardware-implemented modules. Accordingly, the described hardware-implemented modules may be regarded as being communicatively coupled. Where multiple of such hardware-implemented modules exist contemporaneously, communications may be achieved through signal transmission (e.g., over appropriate circuits and buses) that connect the hardware-implemented modules. In embodiments in which multiple hardware-implemented modules are configured or instantiated at different times, communications between such hardware-implemented modules may be achieved, for example, through the storage and retrieval of information in memory structures to which the multiple hardware-implemented modules have access. For example, one hardware-implemented module may perform an operation, and may store the output of that operation in a memory device to which it is communicatively coupled. A further hardware-implemented module may then, at a later time, access the memory device to retrieve and process the stored output. Hardware-implemented modules may also initiate communications with input or output devices.

As illustrated, the computing and networking environment 300 may be a general purpose computing device 300, although it is contemplated that the networking environment 300 may include other computing systems, such as personal computers, server computers, hand-held or laptop devices, tablet devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, digital signal processors, state machines, logic circuitries, distributed computing environments that include any of the above computing systems or devices, and the like.

Components of the general purpose computing device 300 may include various hardware components, such as a processing unit 302, a main memory 304 (e.g., a memory or a system memory), and a system bus 301 that couples various system components of the general purpose computing device 300 to the processing unit 302. The system bus 301 may be any of several types of bus structures including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. For example, such architectures may include Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA (EISA) bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus also known as Mezzanine bus.

The general-purpose computing device 300 may further include a variety of computer-readable media 307 that includes removable/non-removable media and volatile/nonvolatile media but excludes transitory propagated signals. Computer-readable media 307 may also include computer storage media and communication media. Computer storage media includes removable/non-removable media and volatile/nonvolatile media implemented in any method or technology for storage of information, such as computer-readable instructions, data structures, program modules or other data, such as RAM, ROM, EPSOM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that may be used to store the desired information/data and which may be accessed by the general purpose computing device 300. Communication media includes computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. For example, communication media may include wired media such as a wired network or direct-wired connection and wireless media such as acoustic, RF, infrared, and/or other wireless media, or some combination thereof. Computer-readable media may be embodied as a computer program product, such as software stored on computer storage media.

The main memory 304 includes computer storage media in the form of volatile/nonvolatile memory such as read only memory (ROM) and random access memory (RAM). A basic input/output system (BIOS), containing the basic routines that help to transfer information between elements within the general-purpose computing device 300 (e.g., during start-up) is typically stored in ROM. RAM typically contains data and/or program modules that are immediately accessible to and/or presently being operated on by processing unit 302. For example, in one embodiment, data storage 306 holds an operating system, application programs, and other program modules and program data.

Data storage 306 may also include other removable/non-removable, volatile/nonvolatile computer storage media. For example, data storage 306 may be: a hard disk drive that reads from or writes to non-removable, nonvolatile magnetic media; a magnetic disk drive that reads from or writes to a removable, nonvolatile magnetic disk; and/or an optical disk drive that reads from or writes to a removable, nonvolatile optical disk such as a CD-ROM or other optical media. Other removable/non-removable, volatile/nonvolatile computer storage media may include magnetic tape cassettes, flash memory cards, digital versatile disks, digital video tape, solid state RAM, solid state ROM, and the like. The drives and their associated computer storage media provide storage of computer-readable instructions, data structures, program modules and other data for the general purpose computing device 300.

A user may enter commands and information through a user interface 340 or other input devices 345 such as a tablet, electronic digitizer, a microphone, keyboard, and/or pointing device, commonly referred to as mouse, trackball, or touch pad. Other input devices 345 may include a joystick, game pad, satellite dish, scanner, or the like. Additionally, voice inputs, gesture inputs (e.g., via hands or fingers), or other natural user interfaces may also be used with the appropriate input devices, such as a microphone, camera, tablet, touch pad, glove, or other sensor. These and other input devices 345 are often connected to the processing unit 302 through a user interface 340 that is coupled to the system bus 301, but may be connected by other interface and bus structures, such as a parallel port, game port or a universal serial bus (USB). A monitor 360 or other type of display device is also connected to the system bus 301 via user interface 340, such as a video interface. The monitor 360 may also be integrated with a touch-screen panel or the like.

The general purpose computing device 300 may operate in a networked or cloud-computing environment using logical connections of a network Interface 303 to one or more remote devices, such as a remote computer. The remote computer may be a personal computer, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above relative to the general purpose computing device 300. The logical connection may include one or more local area networks (LAN) and one or more wide area networks (WAN), but may also include other networks. Such networking environments are commonplace in offices, enterprise-wide computer networks, intranets and the Internet.

When used in a networked or cloud-computing environment, the general purpose computing device 300 may be connected to a public and/or private network through the network interface 303. In such embodiments, a modem or other means for establishing communications over the network is connected to the system bus 301 via the network interface 303 or other appropriate mechanism. A wireless networking component including an interface and antenna may be coupled through a suitable device such as an access point or peer computer to a network. In a networked environment, program modules depicted relative to the general purpose computing device 300, or portions thereof, may be stored in the remote memory storage device.

It should be understood from the foregoing that, while particular embodiments have been illustrated and described, various modifications can be made thereto without departing from the spirit and scope of the invention as will be apparent to those skilled in the art. Such changes and modifications are within the scope and teachings of this invention as defined in the claims appended hereto.

Claims

1. A system for monitoring hand washing compliance, comprising:

a sensor in communication with a network interface and operable to capture a set of test data;

a computing device in communication with the sensor, the computing device including a processor in communication with a memory, the memory including instructions, which, when executed, cause the processor to: segment the set of test data into a plurality of segments using a change point detection module; align the plurality of segments of the set of test data; extract a feature vector representative of a temporal combination of the plurality of segments; compare the feature vector of the temporal combination of segments against a component model of a compliant routine; obtain an aggregate output representative of a positive detection between the temporal combination of segments and the component model; identify a percentage of positive detections across the plurality of temporal combinations of the plurality of segments; and compare the percentage of positive detections with a threshold percentage to determine whether a corresponding component of the compliant routine is sufficiently present in the set of test data.

2. The system of claim 1, wherein the instructions, when executed, further cause the processor to:

identify a component of the compliant routine that is not sufficiently present in the set of test data by identifying a component model with a lower percentage of positive detection than the threshold percentage.

3. The system of claim 1, wherein the processor is configured to align the plurality of segments of the set of test data using dynamic time warping.

4. The system of claim 1, wherein the processor is configured to represent a handwashing routine in the form of a context-free grammar.

5. The system of claim 1, wherein the sensor is an inertial measurement unit.

6. The system of claim 1, wherein the component model is compared with the feature vector using a support vector machine model.

7. The system of claim 6, wherein the support vector machine model is configured to recognize a component of a plurality of components of the compliant routine.

8. The system of claim 7, wherein an output of the support vector machine model includes a matrix, wherein the matrix denotes whether the support vector machine model recognizes the feature vector as a positive detection of the component of the compliant routine.

9. The system of claim 6, wherein the support vector machine model is trained using a set of training data, wherein the set of training data is of a compliant handwashing routine.

10. The system of claim 1, wherein the instructions, when executed, further cause the processor to:

receive the set of test data from the sensor, wherein the set of test data is of a hand washing routine.

11. The system of claim 1, wherein the instructions, when executed, further cause the processor to:

provide a notification denoting that a component of the compliant routine is not sufficiently present in the test data.

12. A method for monitoring hand washing compliance, comprising:

providing a sensor in communication with a network interface and operable to capture a set of test data and a computing device in communication with the sensor, the computing device including a processor in communication with a memory:

segmenting, by the processor, the set of test data into a plurality of segments using a change point detection module;

aligning the plurality of segments of the set of test data;

extracting a feature vector representative of a temporal combination of the plurality of segments of the test sample;

comparing the feature vector of the temporal combination of segments against a component model of a compliant routine;

obtaining an aggregate output representative of a positive detection between the temporal combination of segments and the component model;

identifying a percentage of positive detections across the plurality of temporal combinations of the plurality of segments of the test sample; and

comparing the percentage of positive detections with a threshold percentage to determine whether a corresponding component of the compliant routine is sufficiently present in the set of test data.

13. The method of claim 12, further comprising

identifying a component of the compliant routine that is not sufficiently present in the set of test data by identifying a component model with a lower percentage of positive detection than the threshold percentage.

14. The method of claim 12, wherein the step of aligning the plurality of segments of the set of test data is performed using dynamic time warping.

15. The system of claim 12, handwashing routine is represented in the form of a context-free grammar.

16. The method of claim 12, wherein the sensor is an inertial measurement unit.

17. The method of claim 12, wherein the component model is compared with the feature vector using a support vector machine model.

18. The method of claim 17, wherein the support vector machine model is configured to recognize a component of a plurality of components of the compliant routine.

19. The method of claim 18, wherein an output of the support vector machine model includes a matrix, wherein the matrix denotes whether the support vector machine model recognizes the feature vector as a positive detection

20. The method of claim 17, wherein the support vector machine model is trained using a set of training data, wherein the set of training data is of a compliant hand washing routine.

21. The method of claim 12, further comprising:

receiving the set of test data from the sensor, wherein the set of test data is of a hand washing routine.

22. The method of claim 12, further comprising:

providing a notification denoting that a component of the compliant routine is not sufficiently present in the set of test data.