
Research & Development Engineer
Advanced Robotic Vision (ARV)
Automated Board Inspection (ABI)
Vitrox Innovation Centre and Headquarter
No. 85-A, Lintang Bayan Lepas 11
Bayan Lepas Industrial Park, Phase 4
11900 Bayan Lepas, Penang, Malaysia
saimun1992 [AT] gmail [DOT] com
(+60)1125427344, (+880)1710277388
saimunurrahman saimunur.rahman
Saimunur_Rahman SaimunurRahman
Hi! I am with the Advanced Robotic Vision (ARV) team of Vitrox Corporation. I lead the machine learning (ML) sub-group in the ARV, Automated Board Inspection (ABI). I have completed my master's (by Research) in Computer Vision and Machine Learning from Visual Processing (ViPr) Laboratory, Faculty of Computing and Informatics, Multimedia University, Cyberjaya Campus under supervision of Dr. John See Su Yang and Dr. Chiung Ching (Peter) Ho. My master degree work was on sptio-temporal feature learning based activity recognition in low quality videos. I was a Graduate Research Assistant and Researcher at ViPr Laboratory, Multimedia University and a Lecturer in Information Technology at International Islamic University Chittagong.
My primary research interests are in designing effective and efficient algorithms for visual recognition tasks in the field of Computer Vision, Pattern Recognition and Machine Learning (both traditional feature learning and deep learning).
News and Highlights
2017
- I am starting a new position with CSIRO Data61 (previously known as NICTA), Australia from March 2018.
- One paper titled 'Exploiting textures for better action recognition in low-quality videos' has been accepted to EURASIP JIVP special issue on 'Applications of Visual Analysis of Human Behaviour'. Link here
- 15.06.2017: I gave a talk on 'Making Machine Understand Beauty: A Photography Perspective' at CoE, Vitrox. Slides here
- Our team developed defect detection system made tremendous progress (from 70% FPY to 98/99% FPY) in V910i (2.5 model) ARV systems.
- 13.04.2017: I gave a talk on 'Beyond the traditional: Learning based defect inspection; A case study on V920i' at CoE, Vitrox. Slides here
- A machine learning enabled defect detection system led by me has been developed and integrated into Vitrox's first ever car engine parts inspection system V920i.
- I have started a new project on Bangladeshi Food Recognition System, few progress including a very new large-scale multi-labelled Bangladeshi food database have been made. Currently, preparation is going on for publications and public disclosure of database. Here is a glimpse of the dataset
- My Master's thesis is out!. Click here to view / download
2016
- I have moved to Vitrox. Here, I will be exploring CVPR opportunies for development of next-generation machine vision inspection systems.
- One paper titled 'Deep CNN Object Features for Improved Action Recognition in Low Quality Videos' presented at ICCSE 2016 (To be published with Adv. Sci. Lett. (SJR Q3)).
- Successfully defended my master's (M.Sc. by Research) thesis!.
2015
- Awarded 'Best Student' for the presentation given in the Visual Computing Research Symposium 2015, MMU.
- Website up and running!
Publications
2017

Human Activity Recognition in Low Quality Videos Using Spatio-Temporal Features
Saimunur Rahman
Master's (by Research) thesis, Multimedia University
June 2016
A dissertation on a generic pipeline and spatio-temporal feature representations for human activity recognition in low quality videos. State-of-the-art results were achieved on nine low quality versions/subsets of publicly available datasets.
PDF BibTex Viva Slides2016

Deep CNN Object Features for Improved Action Recognition in Low Quality Videos
Saimunur Rahman, John See and Chiung Ching Ho
International Conference on Computational Science and Engineering (ICCSE 2016)
November 28-30, Kota Kinabalu, Malaysia, 2016
An image-trained CNN feature utilization method for improving the performance of STIP based traditional HOG and HOF features for action recognition in low quality videos.
PDF BibTex Presentation Slides
Spatio-Temporal Mid-Level Feature Bank for Action Recognition in Low Quality Video
Saimunur Rahman and John See
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2016)
March 20-25, Shanghai, China, 2016
A mid-level feature bank comprises of shape, motion and discriminative textural features for action recognition in low quality videos.
PDF BibTex Poster
Leveraging Textural Features for Recognizing Actions in Low Quality Videos
Saimunur Rahman, John See and Chiung Ching Ho
International Conference on Robotics, Vision, Signal Processing & Power Applications (ROVISP)
February 2-3, Penang, Malaysia, 2016
Proposed BSIF-TOP descriptor for improving shape and motion features for action recognition in low quality videos.
PDF BibTex Presentation Slides2015

On the effects of Low Quality Videos in Human Action Recognition
John See and Saimunur Rahman
International Conference on Digital Image Computing: Techniques and Applications (DICTA)
October 23-25, Adelaide, Australia, 2015
Introduced a LBP utilization method for action recogntion in low quality videos
PDF BibTex Poster
Action Recognition by Jointly using Shape, Motion and Texture Features in Low Quality Videos
Saimunur Rahman, John See and Chiung Ching Ho
IEEE International Conference on Signal and Image Processing Applications (ICSIPA)
October 19-21, Kuala Lampur, Malaysia, 2015
Introduced a nobel joint feature based method for action recogntion in low quality videos
PDF BibTex Presentation Slides2014

An Web-Based Automated Conference Management System
Saimunur Rahman
Bachelor (Engineering) Final Year Project Report, International Islamic University Chittagong
January 2014
A final year project on Web-Based Automated Conference Management System. This project later used for organizing one national-level and one international-level conference.
BibTexTeaching
Spring 2014 (Department of Business Administration, International Islamic University Chittagong)
- Computer Application in Business (New Syllabus, for second semester)
- Computer Application in Business (Old Syllabus, for third semester)
- Management Information System
- Open Source Web Programming
Autumn 2013 (Department of Business Administration, International Islamic University Chittagong)
- Management Information Systems
- Pattren Recognition and Image Processing
- Object Oriented Programming with Java
Downloads
Will be updated soon.