You manage the protocols,we manage from devices to handovers.
The configuration file decides what Pipe reads and how it processes. The template file decides what Cleanse keeps, what each parameter is called, and how it leaves. Between the two, every step runs automatically.
The chain
From devices to handovers
One order, on every trial. The files you manage are the only places the chain asks you a question.
Devices
Whatever the lab captures, and in whatever combination.
Optical
Markerless
Inertial
Force platform
Electromyography
And others
Data
Before any translation.
Data
C3D or a vendor format, carrying the labels your lab happens to use.
Configuration file
You manage it
landmark names
system mapping
processing settings
model version
body measurements
trial hierarchy
Pipe guesses nothing. Without these values there is no anatomical landmark, no model, and no signal setting.
Pipe
Six steps run automatically on every trial, in this order, reading the configuration file at each one.
01
Read and map
Device labels are translated through a common naming convention in Pipe. Sources from different trials can then be treated as equivalent when they are processed.
02
Condition
Filtering, gap filling and downsampling, according to the configuration settings.
03
Calibrate
The skeleton is fitted to this subject: virtual points, segments, body measurements.
04
Reconstruct
The calibrated model is rebuilt frame by frame, for the whole trial.
05
Estimate kinematics
Joint angles and progression angles, for the body regions you asked for.
06
Estimate kinetics
Segment dynamics and normalized joint kinetics projected in the proximal or distal segment.
Custom computations run beside these steps, on the same data and under the same names.
Data
What Pipe hands over.
DataH6D (HDF5)
Angles, moments, powers, envelopes, and more, as continuous series over the whole trial, stored in H6D, an open file format. Nothing has been cut into cycles yet.
Template file
You manage it
parameters of interest
names and aliases
normative reference
export format
how results are laid out
One analytical framework, so parameters measured on different systems can be compared without rework.
Cleanse
Five steps turn continuous series into findings, reading the template file at each one.
07
Verify events
Foot contact and toe off bound every cycle. They are detected on each trial, and any event can be corrected manually.
08
Exclude aberrant cycles
Every cycle is length-normalized on a 0 to 100 % scale and shown per parameter. Discard the unexpected ones from the charts, and the descriptive statistics update as you do it.
09
Compare analyses
Analyses are read side by side, against each other and against the normative reference.
10
Extract features
Spatio-temporal parameters per cycle - cadence, stride length, step time - and summary indices across the whole analysis.
11
Export reports
A clinical report in Word, or a statistics-ready dataset, with one name per parameter and the layout the template describes.
A new parameter becomes available: it is added to the template with its name, and every report after that carries it.
Handovers
Every destination receives the same parameter under the same name.
Report
Patient record
Database
Third-party module
And others
The applications
What each one does
Pipe
Biomechanical processing, run as a batch. You set the configuration once and every trial takes the same path.
Any system, one vocabulary
Optical, markerless, inertial, force platform and electromyography labels are read into shared names.
A fixed sequence for every trial
Conditioning, calibration, reconstruction, kinematics and inverse dynamics, with no manual step between them.
Your own computations
Custom computations run beside the built-in ones, on the same data and under the same names.
An open output
Continuous series stored in H6D, an open file format anything downstream can read.
Cleanse
Motion analysis and the report that comes out of it. The template decides what is kept, what it is called and what is exported.
Event detection and editing
Foot contact and toe off are detected on every trial, and any event can be corrected manually.
Clean cycles, on one scale
Aberrant cycles excluded, the rest length-normalized and averaged against the normative reference.
Features, per cycle and overall
Spatio-temporal parameters computed per cycle, plus summary indices across the whole analysis.
One export, many destinations
A report, a patient record, a database or a third-party module, each receiving the same names.