User experience platform for connected fitness systems

Abstract

Various systems and methods that enhance an exercise or other physical activity performed by a user are described. In some embodiments, a classification system communicates with a media hub to receive images and perform various methods for classifying or detecting poses, exercises, and/or movements performed by a user during an activity. In some embodiments, the systems and methods include a movements database (dB) that stores information as entries relating individual movements to data associated with the individual movements. Various systems, including class generation systems and body focus/activity systems, can utilize the movements database when presenting class content to users and/or presenting exercise information (e.g., muscle groups worked or targeted) to the users.

What is claimed is: 
     
         1 . A connected fitness system, comprising:
 a media hub that captures images of a user performing a workout and presents content to the user via a user interface associated with the media hub;   a classification system that classifies poses or exercises performed by the user from the images captured by the media hub; and   a body focus system that generates content to be presented to the user via the user interface,   wherein the content is generated based on classifications of the poses or exercises performed by the user.   
     
     
         2 . One or more computer memories that store a data structure associated with connected fitness information to be presented to a user of an exercise machine, the data structure including one or more entries, where each of the entries includes:
 information identifying a movement to be performed by a user during an exercise activity; and   metadata associated with the movement to be performed by the user during the exercise activity.   
     
     
         3 . The one or more computer memories of  claim 2 , wherein the movement is a unit of a class presented to the user during the exercise activity. 
     
     
         4 . The one or more computer memories of  claim 2 , wherein the movement is an atomic unit of a class presented to the user during the exercise activity. 
     
     
         5 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes context information for the movement that identifies a body part or muscle group associated with the movement. 
     
     
         6 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes context information for the movement that identifies a description of the movement. 
     
     
         7 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes context information for the movement that identifies an exercise machine or exercise equipment associated with the movement. 
     
     
         8 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes an identifier that represents a machine learning algorithm associated with tracking the movement when the movement is performed by the user during the exercise activity. 
     
     
         9 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes information that identifies related movements. 
     
     
         10 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes information that identifies variations to the movement. 
     
     
         11 . The one or more computer memories of  claim 2 , wherein the metadata associated with the movement to be performed by the user during the exercise activity includes information that identifies content stored in a movement library that is associated with the movement. 
     
     
         12 . A method for presenting workout information to a user performing an exercise activity, the system comprising:
 determining that a user has successfully completed a movement within the exercise activity;   identifying one or more muscle groups associated with the movement; and   presenting information via a user interface associated with the user that represents the identified one or more muscle groups.   
     
     
         13 . The method of  claim 12 , wherein identifying one or more muscle groups associated with the movement includes:
 accessing a movements database that relates movements to metadata associated with the movements; and   extracting, from the metadata associated with the movement successfully completed within the exercise activity, the identified one or more muscle groups associated with the movement.   
     
     
         14 . The method of  claim 12 , wherein presenting information via a user interface associated with the user that represents the identified one or more muscle groups includes presenting a body avatar within the user interface and highlighting, via the body avatar, the one or more muscle groups. 
     
     
         15 . The method of  claim 12 , wherein the user interface is part of a mobile device associated with the user. 
     
     
         16 . The method of  claim 12 , wherein the user interface is part of a display device of an exercise machine utilized by the user during the exercise activity. 
     
     
         17 . The connected fitness system of  claim 1 , wherein the classification system includes:
 a classification network that classifies the poses or exercises performed by the user from the images captured by the media hub; and   a match network that matches the poses or exercises performed by the user during the workout to a template to determine a match prediction for the poses or exercises depicted in images captured by the media hub.   
     
     
         18 . The connected fitness system of  claim 1 , wherein the classification system comprises a machine learning classification network, including:
 a series of encoding layers and decoding layers to generate a predicted keypoint heatmap for the images as a feature map for the images; and   additional downsampling layers and a Softmax function that generate a pose classification or exercise classification from the feature map.   
     
     
         19 . The method of  claim 12 , wherein determining that a user has successfully completed a movement within the exercise activity includes:
 generating, via a neural network, one or more embeddings of images captured of the user performing the movement within the exercise activity; and   predicting the user performed by the movement within the exercise activity by matching, via the neural network, the generated one or more embeddings to a set of template embeddings that represent possible movements performed by the user during the exercise activity.   
     
     
         20 . The method of  claim 12 , wherein determining that a user has successfully completed a movement within the exercise activity includes:
 capturing a sequence of multiple images of a user performing the exercise activity;   identifying one or more inflection points within the sequence of multiple images;   tracking movement of the one or more inflection points within the sequence of multiple images; and   determining the user is performing the movement within the exercise activity based on the tracked movement of the one or more inflection points within the sequence of multiple images.