Meet Ali Hassan, new PhD‑student

Tell me a little about your background?

I’m from Islamabad, Pakistan. I have completed my Master’s degree in Computer Science from COMSATS University, Islamabad, Pakistan. During my Master’s program, I carried out research activities using image processing and deep learning. My Master’s thesis was about deep learning-based region-of-interest (ROI) extraction and ROI-based video compression. After completing my Master’s program, I have gained two years of research experience in deep learning and image processing.

Two years ago, you visited Mid Sweden University on a STINT funded collaboration with COMSATS University of Islamabad (CUI), what have you been doing since then?

I have been on six weeks research visit to Mid Sweden University (Sundsvall) in January-February, 2020. It was a joint project between COMSATS University Islamabad and Mid Sweden University to Investigate Rate-Distortion Optimization for Light Field Compression. One collaborative journal publication, “Computationally Efficient Light Field Image Compression using a Multiview HEVC Framework,” is published in this collaboration. Inspired by this fantastic research experience in the international community, I started looking for Ph.D. opportunities in European Countries. I was surfing the EURAXESS website and found three Early Stage Researcher (ESR) positions at Mid Sweden University. My previous research interest aligned with the project description of the ESR13 position, and I applied for this position. Meanwhile, I was working as Research Assistant at Medical Imaging and Diagnostic Lab, National Centre for Artificial Intelligence, COMSATS University, Islamabad. I was working on identifying and classifying tuberculosis lung diseases (i.e., Computed Tomography Scan, Chest X-Ray, Biopsy/Histopathology) using Deep Learning methodologies in the same lab.

What is it that you are going to investigate as a PhD-student?

My research topic is “Dimension Reduction of Deep Network Models for Mobile Computational Plenoptics.” Deep Learning (DL) architectures are widely used in Light Field (LF) technology and applied to various fields in compression, rendering, and 3D modeling. However, the most successful DL architectures require millions of parameters and operations, making them energy, computation, and memory intensive. Such architectures require long inference time and powerful computational resources (e.g., GPU), resulting in deployment problems on mobile devices having limited hardware resources. Compression of DL architectures is essential for running such architecture on mobile devices. I will be investigating the compression methods for deep learning architecture and the effect of complexity reduction on model performance (i.e., visual quality degradation, output accuracy, etc.).

What have you done so far as a PhD-student at Mid Sweden University?

At the start, I have conducted a literature review on the research topic and identified different compression methods to reduce the complexity of deep learning architecture. Furthermore, I have identified the popular deep learning architectures in Light Field application and tried to compress using the best-identified compression method from the literature study. I have submitted the research findings to an international conference in image processing (ICIP). Besides my research work, I have completed one self-study course related to my research topic.

When did you arrive to Sundsvall, and how has it been so far?

In the last week of September, I came to Sundsvall and joined my responsibilities at Mid Sweden University on October 1, 2021. The winter season was started at the time of arrival. This was my first experience living in a cold city with heavy snow. Moreover, I have witnessed Aurora (Northern Lights) two times yet. Furthermore, I have explored the beauty of Sundsvall by visiting Norra Berget, Södra Berget, and Sidsjön Lake. So far, it has been an outstanding experience in Sundsvall, and I am looking forward to the adventurous journey that awaits me.

Anything more you like to add?

The study and research environments are impressive at Mid Sweden University, and I will get a lot of opportunities to learn and generate new knowledge for the research community. I would like to thank my main supervisor (Prof. Tingting) and co-supervisor (Prof. Mårten and Prof. Karen) for supporting me in leading my research career. I would also like to thank my senior colleague at Mid Sweden University, Mr. Waqas Ahmad, whose support significantly impacted my interpersonal and technical skills.

Read more about the Plenoptima project