updated Feb 27, 2020
first presented at Processing Community Day, 2020 (New Delhi)
As per the Section 144 of the Criminal Procedure Code (CrPC) of 1973 in India the act of "Unlawful Assembly" is "an assembly of five or more persons" for purposes that may not be in the interest of the governing state.
Most recently it was activated in India first during the protests of Shaheen Bagh, New Delhi in early months 2020, and then more rampantly to curb the spread of COVID19 in India.
While experimenting and exploring poseNet, I found that there was a default limitation of the library to recognise only 5 faces at a time. This fits well with "Unlawful Assembly" paradigm that we were going through at the time. This project is not a tool to help state authorities identify "unlawful assemblers" it is more about posing a question and critical look at our excitement about Machine Learning and computer vision algorithms. In the context of current world scenarios where algorithmic bias is becoming an important tool in the hands of the state and machineries of capitalism, it is pertinent to state the right questions with the technology that we use.
You will need to run the project as a localhost on browser.
If you have python3, run the code below in the folder
python3 -m http.server
Aside from a computer with a modern browser you will need a webcam for the project to work.