Thomas Y. Yen

A video-based system for acquiring biomechanical data synchronized with arbitrary events and activities

A video-based data acquisition and interactive multimedia data extraction system are described for measuring and synchronizing large quantities of biomechanical analog data with arbitrary events and activities. Analog signals from up to 32 channels are digitized, frequency-shift key (FSK) coded, and recorded directly onto the audio tracks of a video tape in synchronization with the video information. The data acquisition system includes an A/D converter that digitizes up to 16 multiplexed channels of 8-b data at a fixed sample rate between 60 and 960 Hz, and an FSK modem that transfers the data onto one of two VHS high fidelity (20 Hz-20 kHz bandwidth) audio tracks. Twenty megabytes of digitized data and time codes, along with associated video and normal audio are contained on a conventional 120-min video tape. An analyst interactively reviews the video tape off-line using a computer-controlled VCR and identifies specific events that divide arbitrary activities into time segments. The computer automatically extracts the biomechanical data corresponding to each time segment for further processing or analysis. This system is useful for ergonomics, gait analysis, sports medicine, sleep laboratory, biomechanics, or any application where complex visual events are synchronized with low-frequency analog data.<<ETX>>

Comparison between using spectral analysis of electrogoniometer data and observational analysis to quantify repetitive motion and ergonomic changes in cyclical industrial work

Spectral analysis of continuously measured joint angles using an electrogoniometer was considered as a potentially efficient method for quantifying exposure to physical stress in repetitive manual work. The method was previously demonstrated in the laboratory but has not yet been tested extensively in the field. Spectral analysis was compared against observational analysis, consisting of time-and-motion study and posture classification. Six industrial jobs were selected: (1) press operation, (2) large parts hanging, (3) product packaging, (4) small parts hanging, (5) parts counting and sorting and (6) construction vehicle operation. The posture angle data were synchronized with activities on the video using an interactive multimedia video data acquisition system. Motion for every joint was analyzed using both spectral analysis and observational analysis. Joint angles for the wrist, elbow and shoulder were directly measured using electrogoniometers. Visual posture classification involved determining joint angles from a frozen videotape image sampled three times per s. Repetitiveness was quantified for observational analysis using time study to measure the frequency that specific motions repeat, while spectral analysis measured repetitiveness as the frequency where spectral peaks occurred. Spectral analysis agreed closely with observational analysis. Correlation between the repetition frequencies obtained using time study and spectral analysis was 0.97, with no statistically significant difference observed. Average sustained posture was quantified as the mean, and posture deviation as the RMS angle of joint motion. No statistically significant differences between data obtained using posture classification or spectral analysis were observed for either posture deviation or sustained posture. Since posture classification was very limited in resolution and often contained measurement errors caused by poor joint visibility, the correlation between the postural classification and spectral analysis was 0.77 for sustained posture and 0.53 for posture deviation. When considering only large motions that exceeded the posture classification angle precision, the correlation between postural classification and spectral analysis was 0.81 for sustained posture and 0.81 for posture deviation. Spectral analysis of electrogoniometer data were, therefore, an efficient method for analyzing repetitive manual work that obtained equivalent results, and was more precise than observational analysis.

Exposure assessment of biomechanical stress in repetitive manual work using frequency-weighted filters.

A quantitative exposure assessment strategy for physical stress associated with repetitive manual tasks is proposed using continuous biomechanical data measured directly from electrogoniometers or force sensors. This paper describes an efficient method for reducing large quantities of biomechanical data into a quantifiable metric that accounts for recognized musculoskeletal exposure factors, including repetitiveness, postural or forceful exertion stress, and duration. A frequency domain approach is used for averaging elemental data recorded for repetitive cycles. Parameters for frequency-weighted filters are developed using psychophysical data for equivalent discomfort levels resulting from repetitive movements of different amplitudes and frequencies. These filters enable continuous biomechanical data to be filtered and integrated, resulting in a single quantity corresponding to psychophysical response characteristics for repetitive motion stress. It is anticipated that a similar approach may be used for epidemiological response characteristics. Applications of this theory may make it possible for assessing exposure to physical stress in a manner analogous to the way in which sound level meters are used for measuring exposure to acoustic noise. Repetitive wrist flexion and localized discomfort was used for demonstrating the feasibility of this approach. Suitable data reduction techniques are necessary for evaluating work methods, job designs, and for conducting large scale detailed epidemiological investigations of cumulative trauma disorder risk factors. Frequency-weighted filters based on human response to physical stress at different frequencies can greatly simplify exposure analysis and ultimately may make it possible for quantitative exposure limits to be established.

Automated Video Exposure Assessment of Repetitive Hand Activity Level for a Load Transfer Task

Objective: A new method is described for automatically quantifying repetitive hand activity with the use of digital video processing. Background: The hand activity level (HAL) is widely used for evaluating repetitive hand work. Conventional methods involving either a trained observer on- or off-site or manual off-site video analysis are often considered inaccurate, cumbersome, or impractical for routine work assessment. Method: A cross-correlation-based template-matching algorithm was programmed to track the motion trajectory of a selected region of interest across successive video frames for a single camera to measure repetition frequency, duty cycle, and HAL. A simple, paced, load transfer task was used to simulate a repetitive industrial activity. A total of 12 participants were videoed performing the task for varying HAL conditions. The automatically predicted HAL was compared with the manually measured HAL with the use of frame-by-frame video analysis. Results: Predicted frequency, duty cycle, and HAL were in concert with the manually measured HAL conditions. The linear regression slopes of the automatically predicted values with respect to the manually measured values were 0.98 (R2 = .79), 1.27 (R2 = .63), and 1.06 (R2 = .77) for frequency, duty cycle, and HAL, respectively. Conclusion: A proof-of-concept for automatic video-based direct exposure assessment was demonstrated. Application: The video assessment method for repetitive motion is promising for automatic, unobtrusive, and objective exposure assessment, which may offer broad availability with the use of a camera-enabled mobile device for helping evaluate, prevent, and control exposure to repetitive motions related to upper-extremity injuries in the workplace.

Leveraging Electronic Health Record Systems to Create and Provide Electronic Cancer Survivorship Care Plans: A Pilot Study

PURPOSE The Institute of Medicine (IOM) recommends cancer survivors receive survivorship care plans after completing active cancer treatment. However, care plan creation requires significant time and effort, contributing to diminished adoption of this recommendation. Electronic health record (EHR) systems have been proposed as a solution. We assessed the feasibility of creating and delivering care plans within an EHR system. METHODS Thirty-eight breast cancer survivors without existing care plans were recruited during a follow-up visit to their primary oncologist. Using an EHR template, an oncologist created an individualized care plan for each participant. Time spent creating each plan was recorded. Participant use and feedback were collected. RESULTS Participants enrolled a median of 19.7 months after diagnosis (range, 4.3 to 57 months). A minority of IOM-recommended plan elements could be automatically imported without any manual entry. The majority of elements required interpretation and manual import by the clinician. However, with an established infrastructure for importing elements, the time needed to create a care plan electronically was short (median, 3 minutes; range 2 to 12 minutes). Most survivors (n = 36; 95%) successfully accessed their care plans online and spent a median of 12 minutes (range, 0.5 to 61.9 minutes) reviewing them. Survivors perceived the plans as useful and did not generally report difficulty in accessing them online or understanding content. CONCLUSION Rapid care plan creation and delivery within an EHR is possible. Plans were available to all (survivors, oncologists, primary care physicians) via the EHR. Further research is required to explore the barriers to automating data importation into plans as well as the impact of EHR-integrated plans.

A comparison between analysis time and inter-analyst reliability using spectral analysis of kinematic data and posture classification

This study compares the time needed to analyze data and the inter-analyst variability using observational posture classification vs. spectral analysis of upper limb kinematic measurements made using an electrogoniometer for selected industrial jobs. Eight trained analysts studied four jobs using both methods. An incomplete fixed block experimental design was used, whereby each analyst used one method for each job. The four jobs included (1) punch press operation, (2) packaging, (3) parts hanging, and (4) construction vehicle operation. The posture classification analysis method involved visually classifying tipper extremity joint angles into specific zones relative to the range of motion for every one-third second (10 frames) of videotape. Spectral analysis required the analysts to identify cycle break points. The electrogoniometer signals were synchronized with each cycle, and power spectra for each joint were computed. The average difference in RMS joint deviation among analysts was 0.9 (SD = 0.61 degrees) for spectral analysis and 7.1 (SD = 2.53 degrees) for posture classification. The average difference in mean joint angle was 0.8 (SD = 0.59 degrees) for spectral analysis and 11.4 (SD = 1.58 degrees) for posture classification. Repetition frequency differed an average of 0.05 Hz (SD = 0.054 Hz) for spectral analysis and 0.07 Hz (SD = 0.058 Hz) for posture classification. Posture classification took a factor of 6.3 more time than cycle break point assignment for spectral analysis. Even considering the additional time needed for sensor attachment for direct measurement, posture classification took an average factor of 1.29 more time than spectral analysis using electrogoniometer data.

Use of force plate instrumentation to assess kinetic variables during touch screen use

Touch screens are becoming ubiquitous technology, allowing for enhanced speed and convenience of user interfaces. To date, the majority of touch screen usability studies have focused on timing and accuracy of young, healthy individuals. This information alone may not be sufficient to improve accessibility and usability of touch screens. Kinetic data (e.g. force, impulse, and direction) may provide valuable information regarding human performance during touch screen use. Since kinetic information cannot be measured with a touch screen alone, touch screen-force plate instrumentation, software, and methodology were developed. Individuals with motor control disabilities (Cerebral Palsy and Multiple Sclerosis), as well as gender- and age-matched non-disabled participants, completed a pilot reciprocal tapping task to evaluate the validity of this new instrumentation to quantify touch characteristics. Results indicate that the instrumentation was able to successfully evaluate performance and kinetic characteristics. The kinetic information measured by the new instrumentation provides important insight into touch characteristics which may lead to improved usability and accessibility of touch screens.

Measuring elemental time and duty cycle using automated video processing

Abstract A marker-less 2D video algorithm measured hand kinematics (location, velocity and acceleration) in a paced repetitive laboratory task for varying hand activity levels (HAL). The decision tree (DT) algorithm identified the trajectory of the hand using spatiotemporal relationships during the exertion and rest states. The feature vector training (FVT) method utilised the k-nearest neighbourhood classifier, trained using a set of samples or the first cycle. The average duty cycle (DC) error using the DT algorithm was 2.7%. The FVT algorithm had an average 3.3% error when trained using the first cycle sample of each repetitive task, and had a 2.8% average error when trained using several representative repetitive cycles. Error for HAL was 0.1 for both algorithms, which was considered negligible. Elemental time, stratified by task and subject, were not statistically different from ground truth (p < 0.05). Both algorithms performed well for automatically measuring elapsed time, DC and HAL. Practitioner Summary: A completely automated approach for measuring elapsed time and DC was developed using marker-less video tracking and the tracked kinematic record. Such an approach is automatic, repeatable, objective and unobtrusive, and is suitable for evaluating repetitive exertions, muscle fatigue and manual tasks.

The Accuracy of Conventional 2D Video for Quantifying Upper Limb Kinematics in Repetitive Motion Occupational Tasks

Marker-less 2D video tracking was studied as a practical means to measure upper limb kinematics for ergonomics evaluations. Hand activity level (HAL) can be estimated from speed and duty cycle. Accuracy was measured using a cross-correlation template-matching algorithm for tracking a region of interest on the upper extremities. Ten participants performed a paced load transfer task while varying HAL (2, 4, and 5) and load (2.2 N, 8.9 N and 17.8 N). Speed and acceleration measured from 2D video were compared against ground truth measurements using 3D infrared motion capture. The median absolute difference between 2D video and 3D motion capture was 86.5 mm/s for speed, and 591 mm/s2 for acceleration, and less than 93 mm/s for speed and 656 mm/s2 for acceleration when camera pan and tilt were within ± 30 degrees. Single-camera 2D video had sufficient accuracy ( < 100 mm/s) for evaluating HAL. Practitioner Summary: This study demonstrated that 2D video tracking had sufficient accuracy to measure HAL for ascertaining the American Conference of Government Industrial Hygienists Threshold Limit Value® for repetitive motion when the camera is located within ± 30 degrees off the plane of motion when compared against 3D motion capture for a simulated repetitive motion task.

Force and Impulse Production during the Use of a Touch Screen by Individuals with Motor Control Disabilities

People are increasingly required to interact with touch screens at places ranging from grocery stores to airport kiosks. To date, most of the usability research related to touch screens has included young, healthy subjects. Using novel instrumentation consisting of a force plate and a touch screen, a number entry study examined finger-touch screen interaction by participants with Cerebral Palsy, Multiple Sclerosis, and non-disabled controls. Timing data as well as peak forces and impulses in three dimensions were collected for each touch. The results indicate that, although average peak force vector magnitudes, impulses, and dwell times are similar between the groups, there are significant differences within the same three variables by button size. Average peak force vector magnitude increased by 11 percent while the average vector impulse decreased by 29 percent from the smallest to the largest button size. The average dwell time also decreased 23 percent from the smallest to the largest button size.