INF 2064 – MODVIS: Modeling with Computer Vision

INF 2064 – Topics in Computer Vision, Deep Neural Networks, Geometric Modeling and Rendering

This page is continuously changed as the course progress.

11out19-Sugestão de trabalhos.
08out19-New pptx andd pdfs for topics 05, 06 and 07.
10set19-Conceitos Básicos de Redes Neurais.
10set19-Introdução as Redes Neurais Convolucionais(CNNs).
10set19-YOLO Object Detection (TensorFlow tutorial).
10set19-C4W3L09 YOLO Algorithm (Stanford).
10set19-Lecture 11 | Detection and Segmentation (Stanford).
10set19-An Image Segmentation Tool in JS
01set19-Definition of the T2 Image preparation for Deep Neural Network. Due in: Sep 23.
19ago19-The class moved from the 2nd floor of Pe Laércio Building to the room 511 RDC.
17ago19-Newton's Prism Experiment
17ago19-Printing tutorial: Defining gamut and color space
17ago19-Optical illusions show how we see | Beau Lotto
15ago19-Dr. Craig Blackwell series: Color Vision 1: Color Basics
15ago19-Dr. Craig Blackwell series: Color Vision 2: Color Matching
15ago19-Dr. Craig Blackwell series: Color Vision 3: Color Map
15ago19-Dr. Craig Blackwell series: Color Vision 4: Cones to See Color
15ago19-Interesting experiences: Color Mixer & Accessory Kit
15ago19-Interesting experiences: Colour Mixing: The Mystery of Magenta
15ago19-More on color perception: How do we see color
15ago19-More on color perception: The Science of Color Perception

Professor:

Marcelo Gattass
e-mail: user mgattass at the site [tecgraf or inf].puc-rio.br
Room 6A in the building Pe. Laércio Dias de Moura or room 516 RDC

URL:

http://www.tecgraf.puc-rio.br/~mgattass/modvis/modvis.html

Class Time and Location:

Tuesday 13-15h, Sala 511 RDC (Depto De Informática).

Objective:

The main objective of this course is to detect and geometrically model objects in captured images. These imagens came from different sources. Special emphasis is given for photographic, seismic, MRI, CT and ultra sound images. The course also discusses the geometric model, the mesh representation, camera calibration and rendering algoritms.

Sylabus:

Nature of the data in photographic, medical and seismic images. Classical heuristics and deep neural networks computer vision algorithms to detect and segment features and objects in images. Geometric modeling of objects and polygonal meshes. Rendering and calibration of cameras. Algorithms for virtuam and augumented reality.

Topics:

Colorimetry and images: photographic, seismic, and medical.
Detection and segmentation algorithms based on histograms, neighborhood, graphs, convolutional neural networks (CNNs) and recurrent neural networks (RNN and LSTM).
Geometric modeling of objects and polygonal meshes.
Calibration of cameras and rendering algorithms for extended reality.

Prerequisites:

Basic knowledge of: Linear Algebra, Algorithms, and Phyton programming.

Grading Policy

Assignment #1: (15%) Light spectrum and color reproduction. Due in: Sep 3.
Assignment #2: 15% Image preparation for Deep Neural Network. Due in: Sep 23.
Assignment #3: 15% Rendering by ray tracing and Z-Buffer.
Assignment #4: 15% Camera Calibration and Extended Reality.
Course Project: 40% Free topic related to the course.

Note: Assignments 1, 2, 3 and 4 are developed in groups, but the final assignment is individual. This assignment is a mini thesis: a paper with a presentation that preferably includes a demonstration.

Bibliography:

Maureen C. Stone, A Field Guide to Digital Color, A K Peters, 2003.
Daniel Malacara, Color Vision and Colorimetry - Theory and Applications, SPIE Press, 2002.
Marc Olano, John C. Hart, Wolfgan Heidrich, Michael McCool, Real-Time Shading, A K Peters2002.
Richard Szeliski, Computer Vision: Algorithms and Applications (c) Springuer-Verlag London Limited, 2011.
David A. Forsyth and Jean Ponce, Computer Vision A Modern Approach, Prentice Hall, 2003.
Emanuele Truco and Alessandro Verri, Intoductory Techniques for 3-D Computer Vision, Prentice-Hall, 1998.
Ramesh Jain, Rangachar Kasturi and Bian G. Schunck, Machine Vision, McGraw Hill, 1995.

Important Note:

Nowadays, the web contains lots of excellent material for the course. Please use it. The course also encourages teamwork in assignments 1, 2, 3, and 4. With these three assignments, I hope o create a productive, competitive environment to stimulate discussion and learning. Finally, the last assignment seeks to prepare the individual student for dissertation research.

Here are some usefull links: