CS373 2024 Part3 0 - Introduction

Computer Graphics
and Image Processing

Part 3: Image Processing
Introduction
Patrice Delmas (Lecturer)
Shah (tutor), Tony/Ruigeng (Assignment/Piazza)
1
Images & Computer Vision?
Tzintzuntzan (Michoacan)
2
Ihuatzio (Michoacan)
Lecturer
 Tertiary educated in France (Grenoble)
 Been in New Zealand since 2001
 Intelligent Vision Systems Lab co-founder and director
 Research: Theoretical/Applied Computer Vision at large

 Active projects in Archeology, Marine Sciences,
Behavioral studies, 3D CV. Medical imaging in NZ-
Europe-Mexico-Japan
https://profiles.auckland.ac.nz/p-delmas/publications
3
Credits
 Based on previous years’ lecture slides prepared by
Prof. Georgy Gimel’farb, Assoc Prof Patrice Delmas,
and Dr Martin Urschler.
 Selected Material from:
 Original slides from Image Processing
And Analysis: A Primer by Gimel’farb and
Delmas
 Lecture slides by Richard Alan Peters,
Noah Snavely, Steve Seitz
 Book “Computer Vision: Algorithms and
Applications” by Rick Szeliski
 Book “Digital Image Processing” by
4
R. Gonzalez and R. Woods
What is an image?
5
What is an image?
 Something humans perceive
through their visual system.
light source
object
lens
 Interaction of light sources, geometry
camera
and material properties entering a sensor sensors
(retina, digital camera) 6
What can we do with images?
 Image processing
Sharpened
 Low level operations
Edges
 Image analysis (higher level)
Sky
Mountain
Lake
Grass
8
 Computer vision
 Goal: perceive “story”
behind the image
 Compute properties of
the world
 3D shape
 Recognize people or
objects
 Reason on “What
happened?” (AI)
9
Computer vision
Face detection
 Research:
 PCA for face recognition (Pentland and Turk 1991)
 Robust Face Matching Under Large Occlusions (Gimel’farb, Delmas et al., 2007)
 Applications
Nowadays most cameras detect faces in real time

10
Computer vision
 Login without a password (Biometry)
 Research
 Longitudinal study of fingerprint recognition (Yoon and Jain, 2014)
 On the individuality of fingerprints (Pankanti et al., 2002)
Fingerprint scanners on Face unlock on Apple iPhone X

many new smartphones See also http://www.sensiblevision.com/
and other devices 11
Computer vision
 3D face tracking with
consumer cameras
 Active shape models - their training and application
(Cootes et al., 1995)
 Facial Action Coding System: A Technique for the
Measurement of Facial Movement (Ekman, 1978) Snapchat Lenses
 WG11 (MPEG) (March 2002). "Overview of the MPEG-4 Standard".
12
Face2Face system (Thies et al., 2019)
Computer vision
 Image synthesis (Karras et al., ICLR 2018)
Images synthesized from CelebA-HQ dataset, none of these persons exist! 13

Computer vision
 Smart cars
 Mobileye
 Tesla Autopilot
 Safety features in many high-end cars
14
Computer vision
 Self-driving cars
Google
Waymo
https://www.tesla.com/autopilot
1. Understanding Outliers in Self-Driving

Scenarios (Zhang, Rogers, Delmas,
2022)
2. Enhancing self-driving: Speed bump
https://waymo.com/waymo-driver/ and pothole detection and quantization
15
(Wang, Delmas, 2024)
Back to image processing
16
Back to image processing
 Image processing is a central, fundamental part
of computer vision (see CS 773 to go further in
your understanding of the field or do a project
with us at the IVSLab if you are interested!)
17
COMPSCI 773 contents
Image stitching, Camera calibration, rectification and 3D from stereo matching
(standard and deep learning)
18
IVSLab projects collage
19
IVSLab projects (list)
Medical sciences
 A Framework for the Segmentation of the Cerebral Cortex Laminar Structure
 3D volume segmentation of CT scans using CV-DL hybrid approaches
 SPH Brain deformation
AI for sports
 Deep learning and computer vision approaches for animal reidentification
 Improving temporal consistency of frame-by-frame object detection models applied to
animal data
 3D Localisation of Players in Live Sports Broadcasts
 Person re-identification for sports applications
 Object detection with insufficient labeled dataset by semi-supervised learning
Stand alone
 Enhancing self-driving: Speed bump and pothole detection and quantization
 Integrating AR/VR capabilities to the IVSLab portable multi-camera system
 Quantum Annealing in Computer Vision
Marine Science
 MSI Mussel preservation
 Mapping and understanding the seafloor using AI/CV
20
Compsci 373 part 3 contents
 This part of CS 373 will cover basic image
processing tasks like:
 Globalintensity transformations
 Image filtering, edge detection
 Image segmentation
 Morphological operations
 Assessment
 Quiz/assignment/final exam
21
Assessment
 Quizzes
 Theory and Programming quizzes weekly on week
10-11-12
 The quizzes support the assignment content
 Each quiz is worth 1.25 marks, 7.5 marks in total
 Assignment
A real-life exemplar use of image processing: Coin
separation
 15 marks (10 marks main+5 marks extension)
 Tony/Ruigeng presenting the assignment next
22
What is a digital image?
23
What is a digital image?
A matrix of numbers
24
What is image processing?
 At the root of computer science:
input numbers -> processing -> output numbers
Algorithm
Edge
Detection
25
Let’s do some image processing!
 Lecture recordings
 Python programming examples
 Please prepare a Python3 (>=3.7) environment with matplotlib!
 Coderunner quizzes (Theory and Programming: Python!)
 Image Processing Assignment
“New Zealand Coin detection and identification”
 Ruigeng and Tony will attend
Friday’s lecture to discuss
the assignment
NZ coin detection 26

CS373 2024 Part3 0 - Introduction

Uploaded by

Copyright:

Available Formats

CS373 2024 Part3 0 - Introduction

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CS373 2024 Part3 0 - Introduction

Uploaded by

Copyright:

Available Formats

Computer Graphics

and Image Processing

 Research: Theoretical/Applied Computer Vision at large

 Low level operations

Nowadays most cameras detect faces in real time

Fingerprint scanners on Face unlock on Apple iPhone X

Images synthesized from CelebA-HQ dataset, none of these persons exist! 13

1. Understanding Outliers in Self-Driving

You might also like