postureco
Skip to main content

Have a Question?

Lateral Lumbar Digitization – Computer Vision CBP Technique

Print
< Back

This tutorial will show you how to digitize the lateral lumbar x-ray with Computer Vision.

The computer vision module requires a valid technical support maintenance agreement and is available only in the CBP Technique version. This module interfaces with our proprietary patent-pending cloud software; an active internet connection and current technical support maintenance agreement are therefore required. If the technical support maintenance agreement is cancelled or expires, access to the computer vision module, PostureRay online database backups, and the browser-based Virtual ViewBox will be discontinued. PostureRay remains fully functional in manual digitization mode, exactly as it operated prior to versions 25 and 26.

Computer vision is solely a time-saving tool that places measurement points in an approximate location. It does not replace clinical judgment. The doctor remains fully responsible for validating every single point before completing an assessment. Just as the doctor would be required to review and correct points placed by an assistant, the doctor must review and, when necessary, drag every computer-vision point into the precise anatomical location. Any mis-marked segment must be corrected by the doctor. Only after the doctor has personally validated and accepted all digitization points may the resulting measurements be incorporated into the patient’s health-care record.

The doctor remains fully responsible for any clinical significance assigned to the documentation, because the doctor alone retains complete control over editing and finalizing every annotation point.

Video Link: https://youtu.be/4VKzUFJKsm4

In this above example we demonstrate proper Computer Vision bounding box alignment Note this step is done after any cropping of your x ray image is performed As noted in the video tutorial you must include the vertebral body of L1 not T12 and also include top of sacrum to inferior S1 and get as close as possible based on their alignment If you make this yellow bounding box too large or include T12 this will introduce error By keeping the bounding box as tight as possible to these landmarks reliability and validity is drastically increased Note you do not include the spinouses in the box T12 nor femur heads After you perform computer vision and check landmark locations manually correct any points which are needed Further you can then manually digitize T12 as well as femur heads

Older Manual Method for digitization tutorial is below.

Video link: http://youtu.be/u4pym0mME-Q

Tags: