Undergraduate develops face mask detector

Facial recognition
(Photo: Pexels)
AMMAN — When COVID-19 measures were first put into place, one of the most often flouted rules involved mask wearing. Now, 19-year-old undergraduate, Rakan Armoush, said he has a solution. اضافة اعلان

The developer created an autonomous and adaptive mask face detector, or facial recognition software, that detects those breaking COVID-19 measures.

Despite regulations and penalties issued through Jordan’s Defence Orders to prevent the spread of COVID, wearing face masks is still not universal in Jordan. 

According to a recent report from the Ministry of Industry, Trade, and Supply, more than 20,000 establishments were shut down for violating regulations from October 25, 2020 to January 29, 2021. And inspection teams fined over 108 people and 130 establishments on May 25, 2021 alone.

Embedded into any “real-time camera or webcam”, Armoush’s face mask detector can determine on its own if a person is wearing a mask or not, he said in an interview with Jordan News. 

The sophomore, who studies data science and artificial intelligence at Princess Sumaya University for Technology (PSUT), said he “built the (tracker) from scratch”.

Utilizing artificial intelligence, machine learning, and deep learning, Armoush “programmed and trained the machine” by having it analyze visual imagery and online data sets.

“Machine learning is similar to teaching a young child,” Armoush said. Through exposure to labeled images of human faces, machines extract knowledge from offered data to eventually become able to make independent decisions and cope with dynamic environments.  

“Other than the data set, I developed the whole network,” the young developer added.   

In addition to potentially suppressing COVID-19 transmissions, the face mask detector could have further applications. For example, the detector could possibly be  used to plug loopholes in smartphones’ face IDs, which often fail to recognize a user when they are wearing a mask, Armoush said. 

The aspiring student also suggested that other features could be added to the innovation. “The system might further (be used to) spot incorrect mask wearing instances; as when the mask hangs off one ear, leaves the nose exposed, or rests under the chin,” Armoush said.

‘Aim for more than required’
This is not the first time Armoush has garnered attention, he finished first place in the 2021 WiDS Datathon Stanford competition nationwide, ranking 41st among 808 teams worldwide. 

Armoush and three other student developed “an AI model that forecasts if someone has a certain type of diabetes called diabetes mellitus,” he said. 

Armoush is also the youngest receiver of Google’s TensorFlow Developer Certificate, among only four certified holders in Jordan so far.

The ambitious programmer, who is also the president of the Data Science Club at PSUT, highlighted the importance of extracurricular activities and encouraged students “to aim for more than required, and always take an extra step forward to stand out from the crowd.” 

Armoush addressed some of the beliefs people may have about AI. The legacy of science fiction has fostered the belief that AI is only about humanoid robotics, which might “turn against humans or even conquer the world,” he said.

The developer’s work, however, demonstrates how AI is not a single dimension but can be used in various advancements to “make the world a better place,” he said.

Read more Lifestyle