Dr. Dinesh Kumar

Position: Assistant Lecturer

Email: dinesh.i.kumar@usp.ac.fj

Phone: O: 3232623/ M: 9254727

 

 

 

Biography:

I received my B.A, PGDip and M.Sc. degrees in Computer Science from the University of the South Pacific, FIJI in 1999, 2002 and 2010, respectively; and the Ph. D degree in Computer Science (in the area of Artificial Intelligence with specialisation in Computer Vision, and with applications to health informatics) from the University of Canberra, AUSTRALIA in 2021. My research work spans the areas of pattern recognition, computer vision and multi-modal information extraction and integration with a focus on brain-inspired cognitive architectures. I have also worked on the area of computer graphics with applications to facial expression generation and animation using Facial Action Coding System. I am also an adjunct Professional Associate with the University of Canberra, Australia.


COURSES Teaching:
Undergraduate Courses:
CS001, CS111, CS112, CS241, CS341, CS400, IS122, IS222, IS226, IS314, IS333,

Postgraduate Course: IS421, IS431, IS434

Research Interest

My current research work is centred around the areas of pattern recognition, computer vision and multi-modal information extraction and integration with a focus on brain-inspired cognitive architectures. The outcomes of my research are applied to improving the health and well-being of people; particularly in the area of melanoma detection using clinical and crowd sourced skin lesion images, and development of assistive feedback devices for the blind using live video feeds.

 

Members of organization

  • Adjunct – Professional Associate, Faculty of Science and Technology, University of Canberra, Australia
  • Human Centered Computing Laboratory (HCC), University of Canberra, Australia
  • Member of Open Research Group, University of the South Pacific Suva, Fiji
  • IEEE Member (Professional) & IEEE Computational Intelligence Society Member
  • Professional Member, South Pacific Computer Society (SPACS) (h.ps://www.usp.ac../spcs/index.htm)

 Awards and Prizes (if)

  • Apr 2011 – Gold Medal and the Fijian Holdings Limited Prize for the most outstanding Master of                          Science thesis (The University of the South Pacific, Suva, Fiji)
  • Apr 2021 –  USP A Rank Publication Award (The University of the South Pacific, Suva, Fiji)
  • Aug 2016 – USP Staff Development Award – Partial Funding for PhD (The University of the South                          Pacific, Suva, Fiji)
  • Jun 2016 –  Scholarship Fee Waiver for International Higher Degree by Research Students                                        (University of Canberra, Canberra, Australia)
  • Jan 1999 – Multi-Ethnic Affairs Scholarship for Tertiary Studies (Republic of Fiji Islands                          Government)

List of Publication:

Journals

  • 2022  Kumar D, Sharma D (2022) Feature map augmentation to improve scale invariance in             convolutional neural networks. Journal of Artificial Intelligence and Soft Computing Research               (under review)
  • 2021        Kumar D, Sharma D, Goecke R (2021) Deepzoom: A novel convolution technique using stacked kernel architecture. Elsevier Pattern Recognition (under review)

 

Conferences

(Peer Reviewed)

  • 2021       Sharma P, Sharma A, Kumar D, Sharma A (2021) A strategic weight refinement maneuver for      convolutional neural networks. In: 2021 The International Joint Conference on Neural                Networks (IJCNN), IEEE, doi: 10.1109/IJCNN52387.2021.9533359
  • 2021       Kumar D, Sharma D (2021) Feature map upscaling to improve scale invariance in                convolutional neural networks. In: Farinella G, Radeva P, Braz J, Bouatouch K (eds)    Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and    Computer Graphics Theory and Applications – (Volume 5), Scitepress, vol 5, pp 113–122, doi:   10.5220/0010246001130122
  • 2020       Alzyadat T, Praet S, Chetty G, Goecke R, Hughes D, Kumar D, Welvaert M, Vlahovich N, Waddington G (2020) Automatic segmentation of achilles tendon tissues using deep convolutional neural network. In: Liu M, Yan P, Lian C, Cao X (eds) Machine Learning in               Medical Imaging, Springer International Publishing, Cham, pp 444–454
  • 2020       Kumar D, Sharma D (2020) Multi-modal information extraction and fusion with convolutional      neural networks. In: 2020 International Joint Conference on Neural Networks (IJCNN), IEEE World Congress on Computational Intelligence (IEEE WCCI), pp 1–9, doi: 10.1109/IJCNN48605.2020. 9206803
  • 2020       Kumar D, Sharma D (2020) Distributed information integration in convolutional neural    networks. In: Proceedings of the 15th International Joint Conference on Computer Vision,      Imaging and Computer Graphics Theory and Applications – Volume 5: VISAPP, SciTePress, pp        491–498, doi: 10.5220/0009150404910498
  • 2020       Kumar D, Sharma D, Goecke R (2020) Feature map augmentation to improve rotation    invariance in convolutional neural networks. In: Blanc-Talon J, Delmas P, Philips W, Popescu D, Scheunders P (eds) Advanced Concepts for Intelligent Vision Systems, Springer              International Publishing, Cham, pp 348–359
  • 2019       Kumar D, Sharma D (2019) Enhanced waters 2d muscle model for facial expression          generation. In: Proceedings of the 14th International Joint Conference on Computer Vision,      Imaging and Computer Graphics Theory and Applications – Volume 1 GRAPP: GRAPP,       INSTICC, SciTePress, pp 262–269, doi: 10.5220/0007379302620269
  • 2016       Kumar D, Vanualailai J (2016) Low bandwidth video streaming using facs, facial expression and animation techniques. In: Proceedings of the 11th Joint Conference on Computer Vision,             Imaging and Computer Graphics Theory and Applications – Volume 1 GRAPP: GRAPP,               (VISIGRAPP 2016), INSTICC, SciTePress, pp 226–235, doi: 10.5220/0005718202240233

 

CONFERENCE PROCEEDINGS – ABSTRACT

  • 2011       Kumar D, Vanualailai J (2011) Water muscle model for facial expression generation: Some improvements. In: International Conference on Mathematics of Date, Allahabad, India
  • 2011       Kumar D, Sharma B, Jokhan A (2011) M-learning: A new technological tool in higher education. In: International Conference on Mathematics of Date, Allahabad, India
  • 2006       Kumar D, Hussein I (2006) Mathematical modeling for facial expressions using facs.
    In: 2006 New Zealand Mathematics Colloquium
  • 2006       Kumar D, Li Z (2006) Implementation of an adaptive two-dimensional mesh refinement method based on the law of mass conservation. In: Hawaii International Conference on Statistics, Mathematics and Related Field, Hawaii

 

Books
Chapters

  • 2020       Kumar D, Sharma D (2020) Deep learning for drawing insights from patient data for diagnosis      and treatment – a case study of skin cancer classification. In: Reddy S (ed) Artificial    Intelligence and Healthcare Delivery: Powerful Applications for Improving Healthcare                Outcomes, Abingdon: Routledge, Taylor and Francis Group
  • 2020       Kumar D, Sharma D (2019) Deep learning in gene expression modeling. In: Balas VE, Roy SS, Sharma D, Samui P (eds) Handbook of Deep Learning Applications, Springer International Publishing, Cham, pp 363–383, doi: 10.1007/978-3-030-11479-4 17

 

Projects
Research projects completed:

2016 – 2021

  • Brain-inspired information fusion models to combine local and global features for image classification
  • Augmentation of feature maps in Convolutional Neural Networks (CNNs)
  • Non-trainable global feature extractor techniques for invariant object detection in CNNs
  • Spatial feature extraction techniques from images using large kernels in CNNs
  • DeepZoom convolution technique in CNNs

2016 – 2017

  • Feature selection for Deep Learning algorithms

2006 – 2009

  • Low-bandwidth teleconferencing system using physics-based facial model
  • Using Facial Action Coding System (FACS) to generate facial expressions

 

Industry projects -completed:

  • 2013       The Institute of Internal Auditors (IIA) Fiji Website
  • 2009       Power Systems Control and Monitoring System for Clay Engineering Ltd, Suva, Fiji
  • 2009       Stock Management System (Jewellery products) for Anita Jewellers Ltd, Suva, Fiji
  • 2006       Design and Implementation of Spare Parts Retail Management Information Systems         (SPRMIS) for Vijay Auto Spares Ltd Value, Suva, Fiji
  • 2006       Operations Reporting System for Goodman Fielder International (Fiji) Limited, Suva, Fiji

 

Other(s)
PREPRINTS

  • 2020       Sharma A, Kumar D (2020) Classification with 2-d convolutional neural networks for breast           cancer diagnosis. arXiv preprint arXiv:200703218
  • 2020       Sharma A, Kumar D (2020) Non-image data classification with convolutional neural           networks. arXiv preprint arXiv:200703218v1

THESIS

  • 2020       Kumar D (2020) Multi-modal information extraction and fusion with convolutional neural              networks for classification of scaled images. PhD thesis, University of Canberra, Canberra,      Australia
  • 2010       Kumar D (2010) Animating facial expressions with facial action coding system. Master’s thesis, University of the South Pacific, Suva, Fiji

 

Research Supervision
DIRECTED RESEARCH PROJECT (DRP)

  • 2022       Gurmeet Singh (Principal Supervisor), University of the South Pacific Suva, Fiji
    Title:
    Strategies, risks and challenges of effective Enterprise Resource Planning adoption for       Small, Medium Enterprises (SMEs) in retail sector in Fiji.
  • 2021       Akhilesh Chand Shiuram (co-supervisor), University of the South Pacific Suva, Fiji
    Title: Video Highlight Generation Using Multimodal Machine Learning Approach for     E-sport   Videos (League of Legends)

MASTER

  • 2022       Abhishek Swamy (Principal Supervisor) University of the South Pacific Suva, Fiji
    Title: Detecting Phishing Attacks using Computer Vision Methods
  • 2021       Patrick Sharma (co-supervising), University of the South Pacific Suva, Fiji
      Title: A Strategic Weight Refinement Maneuver for Convolutional Neural Networks