Human gait is a unique signature that provides deep insights into the neurological aspects of the person. Gait analysis can provide us a with options for early detection of various neurological conditions.

The session talks about using Machine Learning frameworks with Random Forest (RF) and Support Vector Machine (SVM) classifiers to train a model for early detection of neurological and neuromuscular diseases using the gait analysis from a robotic gait assistance device.


Outline/Structure of the Talk

  • Introduction
  • Overview of human gait trajectory
  • Identifying gait exception parameters
  • Building training models for classifiers
  • Comparing RF and SVM classifiers
  • Deploying and using the models in the real world
  • Summary

Learning Outcome

Attendees will learn how the different classification models are used for disease detection using gait analysis techniques. They will also learn the challenges in building deploying these training models in the real world.

Target Audience

Data Scientists, ML Engineers, Medical ML enthsiasts

Prerequisites for Attendees

It would be good for attendees to have a basic understanding of human gait analysis techniques in order to appreciate the complexities involved in using it for disease detection.

schedule Submitted 5 months ago

Public Feedback

comment Suggest improvements to the Speaker
  • Anoop Kulkarni
    By Anoop Kulkarni  ~  4 months ago
    reply Reply

    Thanks for the proposal. Which dataset you would be using for training? What issues you see in using your datasets? Could you add these as well to the proposal? 



    • Praveen Srivatsa
      By Praveen Srivatsa  ~  1 month ago
      reply Reply

      I work with Bionic Yantra which has a device for rehabilitation post various injuries including Stroke, SPI etc. We are collecting the datasets on the gait patterns for different injuries. One of the things we found was that we can also use the same for diagnosis. So by identifying the gait pattern, we can map it back to a possible neurological condition. This is in the very early stages, but we are in the early stages of verifying this from the clinical data we have collected.

      (Apologies for the late comment. Had a medical issue in the family. Also I only got the email for the session rejection and not for the comments. So by the time I looked it I assumed it was too late to comment. Look forward to a great event and hope to participate next year.)