I'm from Pakistan and currently working as a Marie Curie Early Stage Researcher for EVOCATION ITN in Visual Computing Lab which is a part of the prestigious ISTI-CNR research institute located in Pisa, Italy. As part of my scholarship, I am also pursuing my PhD studies at University of Pisa. My research interest lies in Machine learning and Computer Vision. I have published two research papers which can be seen on my Google Scholar Profile.
March, 2021 - Present
PhD in Computer Science
Location: Pisa, Italy
Thesis Topic : Multimodal and flexible 3D acquisition
Thesis Supervisor : Dr. Paolo Cignoni
Aug, 2018 - July, 2020
Master's in Computer Science
Location : Kazan, Russia
Thesis topic : Deep learning-based trajectory estimation of vehicles in crowded and crossroad scenarios
Thesis Supervisor : Dr. Ilya Afanasyev
Sept, 2014 - June, 2018
Bachelor's in Electrical Engineering
Location : Lahore, Pakistan
Thesis Topic : Design and Implementation of 50 kV DC power supply
Thesis Supervisor : Dr. Sidra Farid
March, 2021 - Present
Marie Curie Early Stage Researcher (ESR)
I am recruited for ESR07 position in EVOCATION ITN. My research is focused on deep learning based point clouds registration and 3D reconstruction.
Nov, 2020 - March, 2021
I worked on deep learning based object detection and tracking vehicles in Wide Area Motion Imagery (WAMI). I also co-authored a paper which has been accepted in SIU 2021 conference.
HONOURS AND AWARDS
Awarded the prestigious Marie Curie PhD fellowship for PhD studies
Merit based university scholarship for Master's studies at Innopolis university.
Merit based NICT scholarship for undergraduate studies in Pakistan.
Highly proficient in ENGLISH
Native Language: Urdu/Punjabi
Tools (Github, Ubuntu, Anaconda)
ESR07 EVOCATION ITN : Multi-modal and flexible 3D acquisition
I am working on this project currently and the aim of this project is to explore and propose deep learning based novel architectures for point clouds registration and 3D reconstruction.
The project explores the performance of Mask RCNN Benchmark and YOLOv3 on UA-DETRAC dataset and then applies SORT tracker to visualize the trajectory of vehicles in CCTV camera data stream.
Moving vehicle detection in Wide Area Motion Imagery (WAMI) [code]
This project aims to detect and track very small moving vehicles in high altitude wide area motion images. CenterTrack algorithm was trained on moving vehicles in WPAFB 2009 dataset and results were compared with state of the art. The trained algorithm was found 8 times faster than the existing algorithms while achieving similar accuracy.
Real time human pose estimation and classification [code]
This project exploits the computational efficiency of NVIDIA Jetson TX2 to do a real-time rule based human pose classification on USB camera data stream. OpenPose library was installed on the jetson and human joint locations are used to make rules for human poses like standing, sitting, saying hello or stop.
Deep learning based human action recognition using human skeletons [code]
This project aims to recognize human action by training deep learning models on human skeleton structures. Skeleton structures of all RGB images in Berkeley MHAD dataset are obtained using OpenPose library. Different deep learning models like CNN, LSTM and autoencoders are trained on the resultant skeleton structures to learn the spatial and temporal information behind each action.
Face Detection and Recognition [code]
In this project, human faces are detected using a YOLO trained face detection model and then a 128 dimensional face encoding is obtained for every detected face. This encoding is compared with the database of face encodings to recognize the human.
This project aims to exploit the remote hardware capabilities to facilitate computationally expensive tasks on the field. A desktop PC acts a smart space and is connected to a low-end mobile robot using ROS.