Jameel Hassan

I am graduate research student at MBZUAI, Abu Dhabi, UAE in Computer Vision. I am interested in research related to core problems in deep learning, machine learning and computer vision. I am currently working on vision-language multimodal learning and self-supervised learning for better generalization in image understanding under the supervision of Prof. Salman Khan, Prof. Fahad Khan and Dr. Muzammal Naseer as part of IVAL .

I spend my leisure time reading books, writing (Medium blog) and doing Fretwork. See my hobby works here. I enjoy playing both outdoor sports like football and basketball, and board games.

Email  /  Github  /  Google Scholar  /  Linkedin  /  Resume  /  Academic CV

• [Aug 2022] I have started my Masters in Computer Vision at MBZUAI, UAE.
• [Oct 2021] I have joined Veracity AI as a Machine Learning Engineer (Computer Vision).
• [Oct 2021] Our paper in improving solar PV penetration in LV networks by optimization of PV inverter powers has been published in Energies, MDPI journal (IF 3.0).
• [Oct 2021] Our paper Hands Off has been accepted for publication in the 16th IEEE ICIIS conference.
• [Nov 2020] Our preprint on the potential of AI for COVID-19 is on medRxiv.
• [Oct 2020] Our paper Modified Auto Regressive Technique for Univariate Time Series Prediction of Solar Irradiance accepted for publication at 15th IEEE ICIIS conference.
• [Aug 2020] I joined UOP DEEE as a research assistant.
• [Jul 2020] Completed my Bachelor's degree at the Faculty of Engineering, University of Peradeniya.
• [Feb 2019] I will be working at VEGA Innovations as an Engineering intern till May 3rd.

I am primarily interested in underlying problems in deep learning, machine learning and computer vision. Currently I am working on vision-language models and self-supervised learning for better generalization for image understanding.

Previously I have worked on a cross domain research of computer vision and epidemiology to identify social distancing violations such as proximity, interactions and group behaviour using CCTV footages, and on Smart Grids to mitigate voltage violations due to PV penetration in the Low Voltage electricity grid using optimization of PV inverter powers.

Design of a lightweight human pose estimation model, improving the 2 stage stacked hourglass network.

Research on the design of a human-human interaction detection system to identify human interactions and provide a threat level assessment for a given environement, in the context of COVID-19. A computer vision approach is undertaken to identify and localize handshake interactions, mask wearing and group behavior amongst people. This information is then abstracted into a temporal graph and then processed to measure the threat level of the given scene and/or video.

A sensitivity matrix based approach to replace Load flow calculations and a novel two-stage optimization to utilize active and reactive power from PV inverters to mitigate the voltage violation problem in the LV electricity grid due to solar PV penetration.

Detection and localization of handshake interactions from video footage with multi-person settings using deep learning.

Review of the use of AI techniques for spatio-temporal modeling, forecasting, and impact modeling on diverse populations as it relates to COVID-19 and its potential for future applications.

Novel univariate time series prediction of solar irradiance using statistical methods that match the state-of-the-art LSTM neural network method.

Credits to Jon Barron.
Last updated November 2021.