ISMAIL MARZUKI

Ismail M

I’m a Staff Research Engineer (R&I) at InterDigital in Montréal, Canada, where I help shape the next generation of video codecs by designing several novel compression algorithms for publications, standardizations, IPs, and contributions to JVET video coding activities.

My journey started with a B.Eng. in Informatics from UIN Suska Riau (2011). I then earned my M.Eng. (2015) and Ph.D. (2020) in Computer Engineering from Kwangwoon University as part of the Image Processing Systems Lab. Along the way, I worked on making codecs smarter and more efficient, especially in intra/inter prediction optimization, rate control, and perceptual/adaptive quantization for HEVC. I also carried out studies aimed at improving the JEM and VVC ecosystems and contributed to international standardization efforts through JCTVC and MPEG.

I’ve been lucky to pick up experiences across a few great research environments: a research visit at Universiti Teknologi Malaysia (2013) and later a postdoctoral position at KAIST (2021) with the Video & Image Computing Lab. I had the chance to collaborate with the Amazon Prime Video team during my postdoctoral period on perceptual video coding—an exciting mix of real-world constraints, quality science, and codec engineering.

[CV] | [LinkedIn] | [Google Scholar] | [Granted Patents]

Update...

2016, received a scholarship from Kwangwoon University for a Doctoral Degree
2016, my first SCI-E Journal was accepted in the Journal of Real-Time Image Processing
2017, my second journal was accepted in the Journal of Real-Time Image Processing
2017, another journal published in the IEIE Transactions on Smart Processing and Computing
2018, our work was accepted in the IEIE Transactions on Smart Processing and Computing
2020, after a long time, finally published another journal in the IEEE Access
2020, an extension of my previous work in IEEE Access was accepted to the same journal
2020, finally succeeded in defense of my Ph.D thesis. Akhirnya LULUS Ph.D
2021, another extension of my previous work in IEEE Access in 2020 was accepted again to the same journal
2021, my first ever job after Ph.D. Accepted in the BK21 Plus Program for Postdoctoral Researcher, KAIST
2023, start a job at INTERDIGITAL, Canada
2023, finally.... our paper, which I co-authored, was accepted and published in MDPI: Applied Science journal
2026, my mark of 3 years at InterDigital with 16 granted IPs out of 41 filed IP applications from 2023. Way to gooo......

Education

2021 - 2023, Post-Doctoral Researcher at KAIST, South Korea
- Supervisor: Prof. Munchurl Kim
- Project: Deep-learning-based Perceptual Video Coding, a research collaboration with the Amazon Prime Video team
2016 - 2020, Ph.D in Computer Engineering, Kwangwoon University, South Korea
- Supervisor: Prof. Donggyu Sim
- Thesis: A Deep Convolutional Feature-based Rate Control for Random-Access Configuration in HEVC Encoder
2013 - 2015, M.Eng in Computer Engineering, Kwangwoon University, South Korea
- Supervisor: Prof. Donggyu Sim
- Thesis: Tile-level Rate Control for HEVC on Multi-core Platform
2007 - 2011, B.Eng in Informatics, UIN Sultan Syarif Kasim Riau, Indonesia
- Supervisor: Prof. Jasril
- Thesis: Application Design to Embed Text and File into Image and Audio File with Least Significant Bit (LSB) Method

Research Interests

My current and past research topics can be found here, which relate to:

- Video Coding (Compression/ Decompression)
  - Fast Coding
  - Rate Control
  - Quantization Parameter
- Computer Vision
- Deep Learning on Video Coding
- Perceptual Video Coding

Honors and Awards

2009, DIPA Bachelor Scholarship as Outstanding Student
2011, DIPA Support Research Funding for Bachelor
2013-2016, Kwangwoon University, Full Professor Scholarship for Master's Degree
2016-2021, Kwangwoon University, Full Professor Scholarship for Doctoral Degree
2021-2022, BK21 Plus Program for Postdoctoral Researcher, KAIST

Experiences Abroad

University of Technology Malaysia, Johor Bahru, Malaysia (01. 2013 - 07. 2013)
Kwangwoon University, Seoul, South Korea (09. 2013 - 02. 2021)
Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea (03. 2021 - 12. 2022)
InterDigital, Montreal, Quebec, Canada (01. 2023 - Present)

Language

Bahasa Indonesia (Native)
English (Fluent)
Korean (Intermediate)
French (Intermediate)

Publications
Please click on this link to check all other publications

9) Bilateral Attention U-Net with Dissimilarity Attention Gate for Change Detection on Remote Sensing Imageries

Jongseok Lee, Wahyu wiratama, Ismail Marzuki, and Donggyu Sim

MDPI: Applied Science, 2023, doi: https://doi.org/10.3390/app13042485

Convolutional neural network; Deep learning; Remote sensing; Satellite images; Change detection

8) Deep Convolutional Feature-Driven Rate Control for the HEVC Encoders

Ismail Marzuki, Jongseok Lee, Wahyu Wiratama, and Donggyu Sim

IEEE Access, 2021, doi: 10.1109/ACCESS.2021.3132294.

Deep neural network, High efficiency video coding (HEVC), Perceptual video coding, Rate control, Video coding

7) Optimal CTU-Level Rate Control Model for HEVC Based on Deep Convolutional Features

Ismail Marzuki, Jongseok Lee, and Donggyu Sim

IEEE Access, 2020, doi: 10.1109/ACCESS.2020.3022408.

Deep neural network, High efficiency video coding (HEVC), Rate control, Video coding

6) Perceptual Adaptive Quantization Parameter Selection using Deep Convolutional Features for HEVC

Ismail Marzuki and Donggyu Sim

IEEE Access, 2020, doi: 10.1109/ACCESS.2020.2976142

Adaptive quantization parameter, Deep neural network, High efficiency video coding (HEVC), Perceptual quantization parameter, VGG-16~network, Video coding