Notes

These notes accompany the University of Maryland CS class CMSC426: Computer Vision Spring 2019. Thank you to Nitin Sanket and Chahat Deep Singh for creating Project 1, Project 2, and the the Descrption of Rotobrush and SfM. The logistics for the course are available here. Slides from lecture are posted here.

Module 0: Preparation

MATLAB Tutorial

Math Tutorial

Module 1: Color Segmentation using GMM

Project 1: Color Segmentation using GMM

color segmentation, bayer filter, image acquisition, color-space

Module 2: Panaroma Stitching

Project 2: Stitching multiple images seemlessly to create a panorama

Part I: Learning the basics
convolution, filtering, features, edges, corners, SIFT, camera model, pinhole model, calibration, projective geometry, homography

Part II: Panorama Stitching
ANMS, feature descriptor, feature correspondence, homography, ransac, cylinderical projection, blending images,

Module 3: Graph-cut Segmentation

Project 3: Segmentation with Fixations

Description of a Segmentation Algorithm: Rotobrush

Rotobrush

local classifiers, color confidence, shape confidence, local boundary deformation

SfM or SLAM

Explaining Structure from Motion (SfM) or Simultaneous Localization and Mapping (SLAM)

Part I: The Traditional Approach
pinhole model, epipolar geometry, triangulation, PnP, bundle adjustment

Part II: The Modern Approach
factor graphs, GTSAM

Fall 2019 Assignments

Homework 1: Linear Least Squares

Release Date: Jan 31, 2019
Deadline: Feb 7, 2019

Project 1: Color Segmentation using GMM

Release Date: Feb 7, 2019
Deadline: Feb 19, 2019

Homework 2: Image Features and Warping

Deadline: Feb 28, 2019

Project 2: Panorama Stitching

Deadline: March 12, 2019

In-class Midterm

March 14, 2018

Project 3: Segmentation

Deadline: April 11, 2019

Recognition

Deadline: May 16, 2019