DMET 901 Computer Vision

Course Information

Abstract

  • This course presents the field of computer vision from a computer science viewpoint. The basic concepts of image processing should be addressed as many of these concepts are essential to devise good vision algorithms.

Outline

    • Types of images
    • Image arithmetic: addition, subtraction, multiplication, division, blending
    • Geometric operations: scaling, rotation, reflection, translation
    • Image enhancement: histogram, histogram equalization, contrast stretching
    • Digital filters: noise, noise reduction with smoothing filters, mean filter, median filter, Gaussian filter, etc.
    • Feature detectors: edge detectors (Roberts, Sobel, Compass, Canny, Zero Crossing), line detectors, corner detectors (Harris, SUSAN), junction detectors.
    • Image analysis: classification, labelling, segmentation, pyramidal architecture
    • Morphology: dilation, erosion, opening, closing
    • Projective geometry: points, lines, planes, directions, homogeneous coordinates, camera model, projection matrix
    • Stereo vision: disparity estimation, epipolar geometry, fundamental matrix, essential matrix, homography matrix, affine matrix, 3D transformation
    • Point matching: calibrated versus uncalibrated matching, short versus wide baseline matching, correlation techniques (ASD, SAD, VNC), Random Sample Consensus
    • 3D reconstruction: types of 3D reconstruction, 3D point reconstruction and triangulation, 3D line reconstruction, voxelization
    • Motion, object recognition and tracking
       

Objectives

  • The goal of this course is to gain a perspective of the current computer vision topics to help the student prepare for advanced research studies in this field.

RenewSession