Transcript WHAT I LEARNED

REU WEEKS 1 & 2
KRISTIN LAM
MATERIAL COVERED
• MATLAB basics
• Edge Detection
• Harris Corner Detector
• Adaboost Face Detection
• Optical Flow
• Lucas-Kanade Method
• Bag of Features
EDGE DETECTION
• Finds the boundaries of objects
• Edges are defined by the discontinuity of
intensities in the image
• Common Edge Detectors
• Canny
• Sobel
• Laplacian of Gaussian
CANNY
SOBEL
LAPLACIAN OF GAUSSIAN
EDGE DETECTION ASSIGNMENT
FINDING THE GRADIENT
Calculate the gradient of the image.
THRESHOLDING
Pick a threshold and binarize it to get edges that
describe the image well.
PYRAMIDS
Create a pyramid of the image with 3 levels.
HARRIS CORNER DETECTOR
• In addition to edges, corners are used to find
matching points between different frames.
• The corner represents the point where two edges
change.
• The gradient of the image in both directions will
have a high variation, which can be detected.
HARRIS CORNER DETECTOR
ADABOOST FACE DETECTION
OPTICAL FLOW
• Pattern of apparent motion of objects, surfaces,
and edges caused by the relative motion between
the observer and the scene
• Can be used to measure velocities of objects
• Lucas-Kanade method
• Assumes displacement of the object between two frames is
small
• Uses “Least Squares,” a statistical method that limits the
distance between a function and the data points of that
function
OPTICAL FLOW
LUCAS-KANADE
BAG OF FEATURES
• Technique for the visual classification of objects and
categories of objects/textures
POTENTIAL PROJECTS
• Multi-Target Tracking with Social
Behavior Model by Yicong Tian
• Multimodal Data Analysis for the
Detection of Attention Deficit
Hyperactive Disorder by Soumyabrata
Dey