Explore advanced computer vision concepts in this 31-minute lecture from the University of Central Florida's CAP5415 course. Dive into corner detection techniques using auto-correlation and mathematical approaches. Learn about the Histogram of Oriented Gradients (HOG) and gain a comprehensive understanding of the Scale Invariant Feature Transform (SIFT) algorithm. Discover the intricacies of automatic scale selection, orientation estimation, and descriptor formation in SIFT. Examine alternative kernels and local maxima detection in position-scale space. Conclude with a review of local descriptors, enhancing your knowledge of feature extraction and image analysis techniques.
Overview
Syllabus
Intro
Corner Detection by Auto-correlation
Corner Detection: Mathematics The quadratic approximation simplifies to
Histogram of Oriented Gradients
Scale Invariant Feature Transform (SIFT)
Overall Procedure at a High Level
Automatic Scale Selection . Function responses for increasing scale (scale signature)
What Is A Useful Signature Function f?
Alternative kernel
Find local maxima in position-scale space of Dog
SIFT Orientation estimation
SIFT Orientation Normalization
SIFT descriptor formation
SIFT Descriptor Extraction
Review: Local Descriptors
Taught by
UCF CRCV
