View source for 18524615. System And Method For Exercise Type Recognition Using Wearables simplified abstract (Google LLC)

=System And Method For Exercise Type Recognition Using Wearables=


==Organization Name==

[[:Category:Google LLC|Google LLC]]


[[Category:Google LLC]]

==Inventor(s)==

[[:Category:Mark Fralick of San Diego CA (US)|Mark Fralick of San Diego CA (US)]][[Category:Mark Fralick of San Diego CA (US)]]

[[:Category:Brian Chen of Sunnyvale CA (US)|Brian Chen of Sunnyvale CA (US)]][[Category:Brian Chen of Sunnyvale CA (US)]]

==System And Method For Exercise Type Recognition Using Wearables - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18524615 titled 'System And Method For Exercise Type Recognition Using Wearables

==Simplified Explanation==

The present disclosure involves using multiple inertial measurement units (IMUs) to recognize specific user activities, such as different types of exercises and repetitions of those exercises. The IMUs can be found in consumer products like smartwatches and earbuds. Each IMU contains an accelerometer and a gyroscope, each with three axes of measurement, totaling 12 raw measurement streams. A training image consists of multiple subplots or tiles, each representing a separate data stream. This training image is then used to train a machine learning model to identify IMU data as corresponding to a specific type of exercise.

* Inertial measurement units (IMUs) are used to identify user activities like exercises.
* IMUs are present in consumer products such as smartwatches and earbuds.
* Each IMU includes an accelerometer and a gyroscope with three axes of measurement.
* A training image with subplots is used to train a machine learning model to recognize exercise types.

== Potential Applications ==
This technology could be applied in:
* Fitness tracking devices
* Physical therapy monitoring systems
* Sports performance analysis tools

== Problems Solved ==
* Accurately tracking and identifying specific exercises
* Providing real-time feedback on exercise form and technique

== Benefits ==
* Improved exercise tracking and monitoring
* Enhanced user experience with personalized feedback
* Potential for injury prevention through proper exercise form

== Potential Commercial Applications ==
* Fitness wearable companies
* Healthcare technology providers
* Sports training equipment manufacturers

== Possible Prior Art ==
One possible prior art could be the use of IMUs in wearable devices for activity tracking and motion sensing. Another could be machine learning models trained on sensor data to recognize specific activities or gestures.

=== Unanswered Questions ===

=== How does this technology handle variations in individual exercise techniques and forms? ===
The article does not specify how the machine learning model accounts for variations in exercise techniques among different users.

=== What is the accuracy rate of the machine learning model in recognizing different types of exercises? ===
The article does not mention the accuracy rate of the machine learning model in correctly identifying various exercises based on IMU data.


==Original Abstract Submitted==

The present disclosure provides for using multiple inertial measurement units (IMUs) to recognize particular user activity, such as particular types of exercises and repetitions of such exercises. The IMUs may be located in consumer products, such as smartwatches and earbuds. Each IMU may include an accelerometer and a gyroscope, each with three axes of measurement, for a total of 12 raw measurement streams. A training image includes a plurality of subplots or tiles, each depicting a separate data stream. The training image is then used to train a machine learning model to recognize IMU data as corresponding to a particular type of exercise.



[[Category:G06V40/20]]
[[Category:G06V10/778]]
[[Category:G06V10/80]]