AllAboutThatPose

This project was bootstrapped with Create React App.

Inspiration

AllAboutThatPose wants to enable people to master complex movements without the help of a professional trainer. We also want to help people master movements as fast as possible by giving them the best possible advice.

Our special focus is on helping men and women with disabilities, for example blind people.

What it does

AllAboutThatPose records movements (for example Push-Ups) as a sequence of different poses. The observed poses are then analyzed with modern computer vision and deep learning technology to offer the user valuable feedback for improvement.

Blind users can interact with the app through text-to-speech and speech recognition technology and do thus not depend on external help.

How we built it

We used complex, state-of-the-art, deep learning algorithms for computer vision. To make the application more accessible we used modern text-to-speech and speech recognition frameworks. To connect these different technologies into one finished product we used javascript with react.

Challenges we ran into

The challenge was not only in applying the state-of-the-art technology on new problems but also at the intersection of the different technologies we used.

Accomplishments that we are proud of

That we managed to combine many different complex technologies into one product that works well. Posenet is the part that calculates the pose and our own algorithm how good the movement is.

What we learned

Valuable insight in state-of-the-art computer vision techniques.

What's next for AllAboutThatPose

Expand AllAboutThatPose to be able to analyze movements from all areas of life, for example:

• Sports: Deadlifts, long-jump, etc.
• Dances
• In the workplace, to teach employees new manual tasks

Built With

With web technologies, deep learning and computer vision techniques.

Built With

• computer-vision
• deep-learning
• javascript
• love
• react
• speech-recognition
• tensorflow
• text-to-speech