Course Code: AIB553
Enquire

Synopsis

Computer vision is a scientific field that enables computers to “see” –-- understand the content of images and videos and use that information to solve real-world problems without human assistance. This course AIB553 Computer Vision & Applications aims to teach the fundamental concepts and various applications of computer vision. The topics covered include image representation, feature detection and matching, camera model, Convolutional Neural Network (CNN), image classification, face recognition, semantic segmentation and Generative Adversarial Network (GAN).
Level: 5
Credit Units: 5
Presentation Pattern: EVERY REGULAR SEMESTER

Topics

  • Introduction to computer vision
  • Image formation
  • Camera model
  • Feature detection and matching
  • Foundations of Convolutional Neural Networks (CNN)
  • Deep Convolutional Models
  • Image classification
  • Object detection and face recognition
  • Object tracking
  • Image segmentation
  • Generative Adversarial Network (GAN)
  • Computer vision for business

Learning Outcome

  • Appraise the image processing fundamentals
  • Construct robust image matching and stitching
  • Evaluate camera and projection models
  • Critique the fundamental theory and techniques of CNN
  • Formulate a CNN model to solve image classification problem
  • Propose various computer vision applications for business
Part-time Programmes / CET Courses


Who Should Attend

Managers and executives engaged in leveraging advanced Artificial intelligence (AI) and Machine learning (ML) methods for data-driven decision-making to address business challenges.

 

Relevance of Course to employment/upskilling/reskilling

The course addresses crucial competencies by providing students with comprehensive knowledge of computer vision fundamentals and their applications. It equips learners to appraise image processing fundamentals, create robust image matching and stitching solutions, evaluate camera and projection models, and critically analyse Convolutional Neural Network (CNN) theory and techniques. These skills are highly valuable for employment development and job upgrading, as they enable individuals to formulate CNN models for image classification and propose various computer vision applications for businesses. Given the growing importance of computer vision in industries like healthcare, automotive, and retail, this course aligns perfectly with industry needs, preparing graduates to excel in the competitive job market.


Admissions pre-requisite

  • An undergraduate degree or an equivalent qualification from a recognised institution
  • Specific courses may have additional requirements or pre-requisite (e.g. counselling courses). For more information, you may contact the Head of Programme regarding additional course requirements or pre-requisite

Subject to your eligibility and the approval of the Head of Programme, credits earned (up to a cap of 30 credit units) from the completion of Graduate CET Modular courses from the suite of Graduate Programmes may be recognised when admitted to the relevant Graduate Programmes.

 

Schedule

TimeAgenda
Week 1
19:00Course introduction
  • Scope and topics
  • Course schedule
  • Assessments
  • GBA and grouping exercise
20:00Introduction to CV
  • What is CV?
  • Why study CV?
  • CV applications in business
  • Why CV is hard?
21:30Hands-on session
  • Set up your Google Colab
Q&A
Week 2
19:00Image formation
  • Pinhole Camera
  • Focal length, aperture, exposure time
  • Image formation using lens
  • Radial distortion and how to calibrate a camera
  • Typical colour imaging pipeline for digital camera
  • Understand perspective and magnification in image formation
  • Image magnification, vanishing point, forced perspective, perspective distortion
20:15Images & image representations
  • Images
  • Image representation
  • Image transformation: geometric transformation and photometric transformation
21:30Hands-on session
  • Image processing
Q&A
Week 3
19:00Feature Detection
  • Why detect features?
  • Applications using feature detection
  • What are good features
  • How to detect features?
    • Canny Edge Detector, Harris Corner Detector, SIFT
  • How to use detected features
    • Feature matching
    • Image warping and blending
    • Application: create panorama
20:30Image Classification
  • How to train, test and evaluate a classifier
  • Nearest neighbourhood classifier
  • Linear classifier
  • Neural network
Q&A
Week 4
19:00Deep dive to CNN
  • CNN history and classical CNN structures
  • CNN layer types: Fully-connected layer, Convolution layer, Pooling layer
20:45Training and transfer learning
21:30Hands-on session
  • Pretrained image classifier
Q&A
Week 5
19:00Object detection
  • What is OD?
  • Applications of OD
  • Available OD datasets
  • Why OD is hard?
  • R-CNN, Fast R-CNN, Faster R-CNN
  • YOLO
20:00Face detection, recognition and verification
  • Face datasets
  • Face recognition pipeline
  • Face recognition models
20:45Image Segmentation
  • Image segmentation vs object detection vs image classification
  • Types of image segmentation
  • Mask R-CNN
  • Unet
21:30Hands-on session
  • Pretrained object detection model
  • OpenCV GrabCut for foreground segmentation
  • Pretrained U-Net
  • Segment Anything by Meta AI
Q&A
Week 6
19:00SOTA generative models
  • Autoencoders
  • Variational Autoencoders
  • GANs and the capability of GANs
  • Diffusion models
  • SOTA generative models DALL-E 2 explained
21:00Summary of the course

Q&A

 

Assessments

The overall course grade is determined by

  • Others, Assignments
  • Quiz, Case study

 

Trainer info

Dr. Cheng Yao is a Principal AI expert at AIQURIS (a TUV SUD venture). She brings with her nearly a decade of invaluable experience in the cybersecurity and AI sectors. Her contributions to the academic realm include the publication of articles and technical reports on the various subjects of adversarial machine learning, trustworthy AI technologies, as well as system security and privacy. These have been featured prominently in some of the most distinguished academic journals and international conferences. Additionally, she is at the forefront of integrating cutting-edge technologies into the industry. At present, she acts as a member from the Singapore Artificial Intelligence Technical Committee in the ISO/IEC JTC 1/SC 42 on the topic of artificial intelligence.

 

Course Completion requirements

  • Participants are required to achieve at least 75% attendance and pass any prescribed examinations/assessments or submit any course/project work (if any) under the course requirement.
  • Participants are required to complete all surveys and feedbacks related to the course
  • The course fees are reviewed annually and may be revised. The University reserves the right to adjust the course fees without prior notice.
  • Singapore University of Social Sciences reserves the right to amend and/or revise the above schedule without prior notice

 

Course Fees, payment and refund policy

 

  International Participants Singapore Citizens (below 40yrs), Permanent Residents Singapore Citizens (40yrs and above) SkillsFuture Mid - Career Enhanced Subsidy1Enhanced Training Support for SMEs2 (Singaporean and PRs)
Course Fees (A) $3,168.00$2,640.00$2,640.00 $2,640.00
SSG Grant (70%) (B)  $1,848.00$1,848.00$1,848.00
Nett Course fees (A) - (B) = (C) $3,168.00 $792.00$792.00 $792.00
9% GST on Nett course fees (D) $285.12 $71.28$71.28 $71.28
Total nett course fees payable including GST (C) + (D) $3,453,12 $863.28$863.28 $863.28
Less additional funding if eligible under various schemes (F) - -$528.00 $528.00
Total nett course fees payable including GST, after additional funding form the various schemes (E) - (f) = (H) $3,453.12 $863.28 $335.28 $335.28
1 Mid-Career Enhanced Subsidy: Singaporeans aged 40 and above may enjoy subsidies up to 90% of the course fees.
2 Enhanced Training Support for SMEs: SME-sponsored employees (Singapore citizens and PRs) aged 21 and above may enjoy subsidies up to 90% of the course fees.

For the various payment mode, please refer here.


For the refund policy, please refer here. 

 

A written request for a refund must be submitted and is subject to approval.

If written notice of withdrawal is given within the cooling off period1 and before the course start date, a full refund of the fees paid less an administrative charge of $110.00 (exclusive of GST) will be given. No refund will be given for withdrawal thereafter.

1 The cooling off period is defined as 7 working days after payment of course fee.

 

For clarification, please contact the SUSS Academy via the following:
Telephone: +65 6248 0263
Email: [email protected]