ECE516 (ECE516H1S): Intelligent Image Processing
Labs and authentic direct mentorship
The most important part of this course is the labs which offer authentic
direct mentorship with a high degree of involvement from the professor and
other leading experts in imaging,
sensing, meta-sensing, and human machine learning.
Our goal for the undergraduates is to help you get into grad school at
MIT (Prof. Mann's alma mater) or Stanford, or to build the skills you need
to found a great startup or be the world's leader in your chosen field,
and for graduate students, finding a great thesis topic.
Lab topics:
- Fourier transform, wavelet transform, and chirplet transform;
- Machine learning for
computer vision: Radar Vision and LEM neural network (world's first
transform with machine learning built-in);
- Biosignals and biosensing.
In this lab we can build an ultrasound system to image the heart.
[Analysis of
Seismocardiographic Signals Using
Polynomial Chirplet Transform...].
- Brain-Computer Interfaces (InteraXon company co-founded by Mann and
his students));
- Fluid User Interfaces: Build a musical physiotherapy machine based on
an array of ultrasonic lock-in amplifiers for phase-coherent sonar;
- See and photograph sound waves, radio waves, and light waves using your
lock-in amplifier.
- Passive vision:
Many courses on computer vision fail to teach the fundamental concepts of what
sensing is and does. We'll begin with fundamental principles by exploring
first a 1-pixel camera and 1-pixel display, quantigraphic (quantifiable)
sensing, and meta-sensing.
- Begin with fundamentals, e.g. 1-pixel camera and display;
- Quantigraphic sensing: Comparametric Equations;
- Self-driving vehicles, sensing, and meta-sensing;
- Phenomenological augmented reality with Metavision;
- Understanding 3 phase motors and electric vehicles;
- Build your own autonomous e-vehicle...
- Complex-Valued Signal Generators
- Build a signal generator that produces a complex-valued output.
You will fundamentally
understand the difference between positive and negative frequencies
and be able to explain that difference to a 5-year old child!
In later labs you will use this signal generator as the
foundation upon which to build autonomous electric vehicles!
- Phase-coherent detection for active computer vision:
- Active vision systems (sonar, radar, lidar):
Build your own extreme broadband lock-in amplifier;
- Build a sonar vision system for the blind;
- Your final project of your own choosing...
Lab schedule:
Lab 1. What is a camera?
.Pinhole camera (effect of aperture size),
.Mathematical models tan(arctan())...,
.Lens=optional part of lab.
(easy to make from household items).
Lab 2. (Chapter 4 of text) Comparametric Equations (HDR = High Dynamic Range)
.Photocell experiment with laptop computer camera, webcam, or the like,
.Comparagrams and comparagraphs,
.Optional: compositing of images; CCRF.
Lab 3. (Chapter 5 of text) Long-exposure photogaphy, CEMENT, Superposimetrics.
Lab 4. (Chapter 6 of text) Orbits (image stitching), 'Vironment maps.
Lab 5. Active vision, radar, sonar, etc.
.SWIM, sonar (audio) with external microphone or speaker.
Optional additional topics (depending on student interest):
- Machine learning, polyphase machine learning, LEM, radar, sonar, lidar
- RGB moveillance;
- Echocardiography;
- Self-driving vehicle;
- 3-phase signal generator for smart cars.
- Metaveillance standards: smart car certification
- Wearable computing and Intelligent Image Processing: smart vision
Course instructor: Prof. Steve Mann
TAs: Zhao Lu and Jacky Lau