Get It Right

The problem Get It Right solves

The past year has been difficult for the entire world in many spheres, be it personal or professional. But the most prominent issue was maintaining our health and lifestyles. We resorted to various forms of home-workouts, which sadly had more adverse consequences than positive results. This was mainly due to insufficient knowledge of warmups and postures, which resulted in performing the exercise wrong. The sudden surge of heavy-duty positions made our bodies dangerously sore and we would wonder if we had accidentally nicked an artery or burst a blood vessel. Keeping this in mind, we have developed a model which can help us assess our exercise postures using Posenet app extension and Mobile net Deep learning model. The app uses video recognition technology to assess the postures and provide us results. The was constructed using the MIT's Personal Image Classifier and the app was constructed over MIT App inventor.

Challenges we ran into

Posenet app extension model used - output wasn't displaying properly while integrating in the app even though the steps to integrate were right. Mobile net model - took time to get the hyper-tuning of the model right, importing the model to the app was a major obstacle

Built With

• deep-learning
• firebase
• machine-learning
• mit-app-inventor
• mit-personal-image-classifier
• mobile-net
• posenet-extension