Cognitive services enable us to recognize patterns and identify characteristics from images, audio or a video. While most of the examples show how we can work with photos and identify pets, scenes or friends, cognitive services also has deep application in key vital areas.

One of the areas that AI and cognitive services can make a big impact is in the area of medical rehabilitation. Understanding the human reaction to a rehabilitation process is vital to assess how their body is reacting to the treatment being provided to them. In this session, we will look at the complete life-cycle of a medical rehabilitation platform and how it uses IoT data, VisionAPI and sentiment analysis to provide a comprehensive feedback on patient rehabilitation to a doctor. It also touches up the sensitive topic of data ownership and how - in this age of data protection - we can use and work with data responsibly.


Outline/Structure of the Case Study

  • Overview of the medical rehabilitation process
  • The challenges that practitioners face today
  • Setting up an IoT Hub and registering IoT devices
  • Using IoT sensors for tracking activity
  • Capturing video and extracting image frames
  • Using sentiment analysis to analyze impact of rehab on patient

Learning Outcome

The attendees will be able to understand a real-life scenario that can improved with cognitive analysis and also understand the process of how we went about enabling the cognitive feature as part of a medical rehabilitation process.

Target Audience

If you are looking at how to leverage cognitive services and vision API in the real world, this session is for you.

Prerequisites for Attendees

Participants should be familiar with general cognitive services like image analysis, recognition etc.

schedule Submitted 2 years ago

  • Praveen Srivatsa

    Praveen Srivatsa - Machine Learning for medical rehabilitation

    Praveen Srivatsa
    Praveen Srivatsa
    Asthrasoft Consulting
    schedule 2 years ago
    Sold Out!
    45 Mins
    Case Study

    Walking is one of the most common human activity. But the human gait varies by gender, age, culture etc. How can we use pre-trained models to identify human gait across different images.

    In this session, we take a look at a real world case study where we are using deep learning models and Vision algorithms like DeeperCut and ArtTrack to objectively measure the human pose and gait and use this as a measure to predict their rehabilitation helping them to get back onto their feet in weeks instead of months.

    We will look at how we went about building and training the model to understand the human gait. We will also look at the challenges that we faced when we wanted to use a generic model to understand Indian patients. We will also touch about the importance of trust and accuracy when working with machine learning algorithms in the healthcare space.