Detect politicians touching smartphones during meetings with machine learning and image recognition

Even though it is in the middle of a meeting, the broadcast may show politicians playing by hand or taking a nap. A mechanism is in operation that automatically detects this from live video using machine learning and image recognition, and reports it with video on Twitter and Instagram accounts.

The Flemish Scrollers, 2021-2022 – Dries Depoorter

Machine Learning Detects Distracted Politicians | Hackaday

Belgian artist Doris Deporter has been running the program ' Flemish Scrollers ' written in Python since July 5, 2021.

The subject of the check is the Flemish Parliament in the Flemish region of Belgium, which broadcasts the state of the parliament on YouTube.

Vlaams Parlement-YouTube

'Flemish Scroller' is a mechanism that detects the appearance of a member who has begun to touch a smartphone during a parliamentary broadcast, cuts out a video of about 7 seconds, and then sends it on Twitter (@FlemishScroller ) and Instagram ( @theflemishscrollers). When there is no live broadcast, the video delivered in the past will be checked.

Since each member of the Diet has an SNS, he skips mentions to himself when posting on Twitter and Instagram, and I'm not sure if touching the smartphone is a legitimate purpose, but at least 'I did not know' It is a mechanism that can not be said.

One example is like this, and it has been found that Prime Minister Jan Jambon of the Flemish government has pulled the smartphone to his hand and started to operate it.

in Software,   Web Service, Posted by logc_nt