Mr Feeham Salam

Research project: Human action recognition from a single video clip


Video recording has become an inseparable part of our daily routine, they include recording of our activities taking place for example in shopping centres, gym, train stations, bars, hotels, banks, universities, hospitals, airports, and holiday places.

The need to automate the process of recognising and understand the activities in video data becomes essential as this requires tremendous amount of manpower to categorise and recognise human actions in videos, for e.g., in surveillance systems (JIN C-B et al., 2017), application includes healthcare systems for injury rehabilitation purposes where a patient's actions can be automatically recognised (Pravin Dhulekar et al., 2017). Many sports required precise actions to be performed in a particular way such as tennis (Ullah et al., 2021). Another application is in entertainment or computer gaming where action recognition is at the core of some gaming consoles like Kinect.

However, many of the above-mentioned systems required specialised hardware/sensors such as 3D cameras or depth-sensors for capturing human action sequences. Consequently, not only does this limit their practicability, but also prevents processing existing video material, e.g., YouTube videos and films produced by the TV and cinema industry. The proposed project aims to address the challenge of efficiently recognise human action from a single video clip captured with either a fixed or moving camera


3 years experience of programming languages like Java, Python and C++.

Web-Development - HTML, CSS, PHP, JavaScript, React and Vue-JS.

Databases - MySQL, ORACLE

First-Class BSc - Kingston University - Computer Science - 2017-2019

Areas of research interest

  • Deep Learning
  • Machine Learning
  • Programming Languages
  • Teaching/Coaching
  • Web Development and Databases


  • BSc - Computer Science

Funding or awards received

  • Faculty of Science, Engineering and Computing Bursary Award – October 2021