Machine Learning Helps You Get In Shape While Working A Desk Job

Humans weren’t created to sit in front of a computer system all working day, nevertheless for a lot of of us that’s how we shell out a large part of our life. Of study course we all know that it’s essential to get up and shift close to just about every now and then to stretch our muscle tissues and get our blood flowing, but it is effortless to ignore if you are doing the job in direction of a deadline. [Victor Sonck] imagined he required some reminders — as effectively as some not-so-light nudging — to get into the behavior of carrying out a brief work out a handful of periods a working day.

To this finish, he built a piece of application that would lock his computer’s screen and only unlock it if he executed five press-ups. Locking the display screen on his Linux box was as quick as sending a command through the network, but recognizing thrust-ups was a harder undertaking for which [Victor] determined to utilize device mastering. A Raspberry Pi with a webcam hooked up could do the trick, but the constrained processing electrical power of the Pi’s CPU may prove insufficient for processing tons of uncooked impression data.

[Victor] as a result made a decision on applying a Luxonis OAK-1, which is a 4K digital camera with a designed-in machine-learning processor. It can run many kinds of image recognition programs together with Blazepose, a pre-educated product that can acknowledge a person’s pose from an impression. The OAK-1 makes use of this to deliver out a established of coordinates that describe the position of a person’s head, torso and limbs to the Raspberry Pi by means of a USB interface. A second equipment-learning model operating on the Pi then analyzes this dataset to understand press-ups.

[Victor]’s video (embedded below) is an entertaining introduction into the planet of device-discovering methods for video processing, as well as a great fingers-on example of a task that results in a beneficial tool. If you are intrigued in mastering much more about machine discovering on compact platforms, check out this 2020 Remoticon talk on machine finding out on microcontrollers, or this 2019 Supercon communicate about implementing equipment eyesight on a Raspberry Pi. out?v=ZiOr9EdYEeE