Today the U.S. Patent and Trademark Office officially granted Apple a patent that relates to devices, systems, and methods that determine a wrist measurement or watch band size using depth data captured by a depth sensor from one or more rotational orientations of the wrist. The new measurement system using Machine Learning is mostly focused on an in-store system.

Touchless Wrist Measurement

Apple's patent application covers devices, systems, and methods that determine a wrist measurement or watch band size using depth data captured by a depth sensor.

In some implementations, wrist measurement techniques are touchless. In some implementations, the wrist measurement techniques are performed by a trained technician in a brick and mortar retail store.

Alternatively, the wrist measurement techniques are performed by an individual using an electronic device like a smartphone that includes the depth sensor. For example, the electronic device is placed on a surface with the depth sensor facing up, and a user may rotate their hand/wrist above the electronic device while at least two depth map images of the wrist are captured during a wrist scanning process.

In some implementations, the depth data includes at least two depth map images of a wrist from different angles sufficiently separated to accurately represent the wrist's circumference.

For example, one of the depth map images of the user's wrist may be captured with the palm facing the depth sensor and the other of the depth map images of the user's wrist is captured with the palm facing to the side (e.g., approximately 90° rotation).

In some implementations, the depth data is input into a machine learning (ML) model that outputs a measurement corresponding to circumference of the wrist and/or a watch band size among a plurality of band sizes. In some implementations, the ML model is a convolutional neural network (CNN) regressor.

In some implementations, at an electronic device having a processor, a method includes obtaining depth data captured by a depth sensor, the depth data including at least two depth map images of a wrist from different angles.

In some implementations, an output is generated based on inputting the depth data into a machine learning model, the output corresponding to circumference of the wrist or a watch band size of the wrist, and a watch band size recommendation is provided based on the output.

In some implementations, the watch band is not adjustable (e.g., no clasp, buckle, notch, etc.), although a non-adjustable watch band may stretch a little to slide over the hand and onto the wrist. An accurate wrist measurement is especially useful for non-adjustable watches.

Apple patent FIGS. 1-5 below illustrate an example of wrist measurement techniques.

To review the full details of this invention, check out granted patent 12223669.   

Apple Inventors

• Aditya Sankar: Senior Research Scientist
• Qi Shan: Apple Intelligence, Vision Pro
• Shreyas Joshi: Senior Machine Learning Engineer
• David Güera: Senior Research Scientist
• Fareeha Irfan: Machine Learning Engineer
• Bryan Perfetti: Computer Vision Algorithm Engineer