Muhamamd Aasim Rafique

I am a faculty memeber in the department of computer science at King Faisal University. My areas of research include particularly machine learning in general artificial intelligence.

I have worked on self-driving vehicles, neuromorphic devices and computer vision during my stay at Machine Learning and Vision lab, where I spent time as a reseach professor, postdoc and a phd student. The title of my phd thesis is "Predictive Learning with Generative Modeling". I see singularity complenting homo sapiens where later being centrifugal force in disturbing inertia.

Email  /  CV  /  Bio  /  Google Scholar  /  Twitter  /  Github  / 
Research statement  /  Teaching statement  /  Genealogy Tree

I'm interested in computer vision, machine learning and image processing and applications in self-driving vehicles. My research includes applications of AI with a focus on understanding of cognition in the universe. I worked with neuromorphic devices, self-driving vehicles, intelligent transportation and general AI in general. Some self-driving vehicles related research papers are highligeted in YELLOW and generative modeling papers are highlighted in GREEN.

DepthSense employs an encoder-decoder architecture, a Radar Residual Network, feature fusion with a spatial attention mechanism, and an ordinal regression layer to deliver precise depth estimations.

Exploiting temporal bias in the videos for calculation of morphological properties during growth of nanopaticles.

Introducing a loss function penalizing the inferrence error of the peculiar spikes in Au Spiky Nanoparticles refines the segmentation and consequntly improves the morphology statistics.

Introducing a loss function penalizing the inferrence error of the peculiar spikes in Au Spiky Nanoparticles refines the segmentation and consequntly improves the morphology statistics.

Introducing additional label attention and objects association in images for improved image caption generation.

Explored hard inductive bias of pyramid features with soft-inductive bias of transformers.

Proposes a domain adaptation framework for transfer learing from RGB to Thermal images.

Proposes a Gaussian Process for occupancy mapping from sparse radar data.

Employed deep neural networks for teaching methodology analysis and using the statistics for improvements.

Used a sequential netwrok to learn from the driver behavior captured with CAN Bus data.

Used a consensus based improvement in dcoument clustering.

Proposed a generative model for pixel based prediction of background segmentation.

A clustering based algorithm to track objects.

A generative technique to propose regions for the license plate and then classify the regions as plates.

A generative model used with memristive weights for speech classification.

A RBM training algorithm designed to be used with real memristor weights.

An architecture for the digital twin of an autonomous vechile.

Proposed an auxilary loss for CAN Bus data prediction which is inspired from the dominating visual cues.

Proposed a DNN and evalated a series of techniques to generate statistics of the morhphology of a nano particle.

Used a single camera for the localization and identification of the markers to follow, with no prioir localization information.

1. S. Azam, F. Munir, Rafique, Muhammad Aasim and M. Jeon, “Object Modeling from 3D Point Cloud Data for Self-Driving Vehicles,“IV’18, Changshu, Suzhou, China, June 26-30, 2018
2. S. Azam, M. Rafique, Muhammad Aasim, A.M. Sheri and M. Jeon, “Mutli-sensor data collection and data fusion: Step towards self-driving cars,“BIC, Jeju Island, 2017.
3. Munir, F., Azam, S., Rafique, Muhammad Aasim, & Jeon, M. (2017). Automated Labeling of 3D Point Cloud Data. Korea Information Science Society KSC, 769-771, 2017
4. S. Azam, Rafique, Muhammad Aasim, and M. Jeon, “Vehicle Pose Detection Using Region Based Convolutional Neural Network", 2016 International Conference on Control, Automation and Information Sciences (ICCAIS), Ansan, OCT- 27-29, 2016, pp. 194-198. doi: 10.1109/ICCAIS.2016.7822459

• Rafique, Muhammad Aasim, Azam Muhammad Shoaib, Choi In-Moon, Jeon Moongu "Face DeIdentification Method", Domestic Application No .: 2017-0032265, Date of Application: 2017.03.15, Registration No .: 10-1861520 (Republic of Korea)
• Rafique, Muhammad Aasim, Abdulmohsen Albasher, AbdulRaheem Ahmad "ADAPTIVE KEYBOARD FOR TYPING CORRECTION", Application No .: 18/542,295 , Date of Application: 2023.12.15 , (Filed US)
• Rafique, Muhammad Aasim, Mohammed Elhassan, Qazi Mudassir Ilyas, "A propeller convertible to a robotic arm", Application No. 18/412,039 Date of Application: 2024.12.1, (Filed US)

• Program Chair and Member Organizing Committee: 2nd DeepView Workshop in Parallel with IEEE AVSS 2022
• Member technical committee 10th ICCMA 2022
• Program Chair and Member Organizing Committee: 1st DeepView Workshop in Parallel with IEEE AVSS 2021
• Sponsorship Co-Chair, IEEE ICoDT2 2021
• Technical Co-Chair, IEEE ICET 2018

I stole the website template from Jon Barron, as he states, "Feel free to steal this website's source code. Do not scrape the HTML from this page itself, as it includes analytics tags that you do not want on your own website — use the github code instead. Also, consider using Leonid Keselman's Jekyll fork".