He walked up the podium cheered on by the graduating class, some of whom had been his students, and he walked off a Doctor of Philosophy in Information Technology with a thesis titled: a concealed firearm detection approach on video using skeletal tracking and supervised machine learning techniques.

In the study, a ceska pistol with a weight of 2.77kg and a length of 0.2 meters was used. This firearm was used because it’s the most illegally used firearm in Kenya. Authorization to use the firearm for the study was granted by the office of the Inspector General of Police in Kenya.

All participants were dressed in trousers and a long jacket or sweater to conceal the firearm. They walked towards the sensor for 4.5 meters in two recorded trials: when they were not carrying a firearm and when carrying the firearm concealed on the right hip.

Some of his students assisted him in his study as they acted as subjects carrying firearms, while Dr. Muchiri used a Kinect v2 camera to capture the 3D skeletal gait features and analyse these features using machine learning techniques with an aim of distinguishing between armed and unarmed persons. The findings indicated the potential of artificial intelligence techniques to detect persons carrying concealed firearms on video using skeletal gait features with over 90% accuracy.

Why did you choose the thesis topic?

I had read extensively about the use of metal detectors and their limitations especially to do with their invasive nature. The gadgets are still effective and quite popular because they are budget friendly and they get the job done. However, some issues arise with metal detectors. For instance, they bring about increased foot traffic at airports, malls, offices and other public places. Moreover, a person carrying a firearm would be detected a bit too late and this poses a security risk. So I thought of a way out. Coincidentally at the time when I was conceptualising the idea, the street cameras in Nairobi had just been launched. This, together with my idea, turned to what is now my doctorate.

How does this system of concealed firearm detection work?

The developed approach uses an RGB-D skeletal tracking camera to track the gait features of people and uses machine learning to analyse these gait features and determine whether a person is carrying a concealed handgun or not. The camera used in the study could detect a person with a firearm from about 4 metres away.

What were some of the challenges you faced when doing your PhD?

Balancing my PhD studies and teaching at the Faculty of Information Technology required a lot of discipline. My PhD project was tough which meant a steep learning curve. I married while pursuing the PhD; I explained my situation to my wife and she was very supportive during this period as I tried to juggle all aspects.

Tell us about your Strathmore experience.

Strathmore is home. I joined Strathmore in 2004 as an undergraduate student in the Bachelor of Business Information Technology programme, and later joined the Master of Science in Information Technology programme. The infrastructure has definitely developed over the years; what is now the Student Centre was the hockey pitch when I first joined the University.

Have you gotten used to people calling you Daktari?

No, it’s a very strange feeling. My father is a doctor by profession so when I hear the Dr. Muchiri, I automatically assume it is him that people are referring to.

What’s next now that you have your PhD?

I look forward to being a professor. I would have to settle in first of course but the hardest part of it all is complete. What I will focus on now is more research on my project, mentoring PhD students, and increasing my number of publications.

What activities do you take part in during your free time?

I like to read, watch plays, and work out though that took a back seat while I was doing my PhD but it’s something I would like to get back into. There’s a book I’m currently reading with my wife, The road less travelled, by Scott Peck.

I am keen on mentoring young men. I would like to get a better understanding of what’s happening to our boys and why some of them are not reaching their full potential. I believe that one way to deal with this issue is by having strong male role models who can be emulated.

What advice do you have for those who are doing their PhD?

Don’t give up. There are times when you will work on something for three weeks to a month only to realise that you’ll have to start afresh. Don’t lose hope. Be patient.

 

This article was written by Wambui Gachari.

If you have a story, kindly email: communications@strathmore.edu