Awais Nizamani

I am an AI Phd Candidate at Murdoch University (MU) and The University of Western Australia (UWA) in Perth Australia, where I am working on Intelligent Virtual Humans under the supervision of Prof. Hamid Laga, Prof. Mohammed Bennamoun, and Prof. Farid Boussaid, which is funded by the Australia Research Council (ARC).

Prior to PhD, I worked at Retrocausal for 3 years as an Applied Research Engineer (Computer Vision) under the supervision of Dr. Zeeshan Zia and Dr. Quoc-Huy Tran.

In 2020, I completed my undergrad Computer Science from FAST-NUCES Karachi, Pakistan, where I worked on Multimodal classification datasets under the guidance of Prof. Tahir Syed.

Email / CV / Scholar / Linkedin / Github

Publications

I am interested in computer vision, deep learning, generative AI, and image processing. My research is about video understanding and shape analysis inferring the shape and motion from 3D data or videos. Papers are highlighted in reverse chronological order. '*' denotes equal contribution.

	Dynamic Neural Surfaces for Elastic 4D Shape Representation and Analysis Awais Nizamani, Hamid Laga, Gaunjin Wang, Farid Boussaid, Mohammed Bennamoun, Anuj Srivastava CVPR*, 2025 project page / arXiv We introduce Dynamic Neural Surfaces (DNS) for 4D shape representation, enabling elastic and diffeomorphic modeling of complex shape dynamics.
	Unsupervised Action Segmentation by Joint Representation Learning and Online Clustering Sateesh Kumar, Sanjay Haresh, Awais Ahmed, Andrey Konin, M. Zeeshan Zia, Quoc-Huy Tran CVPR, 2022 project page / arXiv A novel unsupervised approach for action segmentation, combining representation learning with online clustering for temporal consistency.
	Timestamp Supervised Action Segmentation with Graph Convolutional Networks Hamza Khan, Sanjay Haresh, Awais Ahmed, Shakeeb Siddiqui, Andrey Konin, M. Zeeshan Zia, Quoc-Huy Tran IROS*, 2022 project page / arXiv We propose a GCN-based framework for timestamp-supervised action segmentation, reducing annotation requirements while maintaining accuracy.
	Dataset Augmentation Strategies for Visual Activity Recognition in Deep Neural Networks Awais Ahmed Nizamani ICCC, 2022 paper A study on augmentation techniques for enhancing generalization in visual activity recognition using deep neural networks.
	AI-mediated Job Status Tracking in AR as a No-Code Service Andrey Konin, Shakeeb Siddiqui, Hasan Gilani, Muhammad Mudassir, M. Hassan Ahmed, Taban Shaukat, Muhammad Naufil, Awais Ahmed, Quoc-Huy Tran, M. Zeeshan Zia ISMAR, 2022 project page / paper We present a no-code AI-mediated AR system for job tracking, enabling intuitive monitoring and reporting in industrial workflows.
	System and Method for Determining Sub-activities in Videos and Segmenting with Little to No Annotation Muhammad Shakeeb Hussain Siddiqui, Quoc-Huy Tran, Muhammad Zeeshan Zia, Andrey Konin, Sateesh Kumar, Sanjay Haresh, Awais Ahmed, Hamza Khan US Patent, 2022 patent link Patent describing a system for fine-grained video activity segmentation with minimal supervision.