Lecture-14.ppt - Change Detection, Skin Detection, Color...

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Change Detection, Skin Detection, Color Tracking Lecture-14 Mixture of Gaussians Grimson
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Algorithm • Learn background model by watching 30 second video • Detect moving object by measuring deviations from background model, and applying connected component to foreground pixels. • Predict position of a region in the next frame using Kalman filter • Update background and blob statistics Summary • Each pixel is an independent statistical process, which may be combination of several processes. – Swaying branches of tree result in a bimodal behavior of pixel intensity. • The intensity is fit with a mixture of K Gaussians. ) ( ) ( 2 1 1 2 1 2 1 | | ) 2 ( ) Pr( j t j T j t X X K j j m j t e X µ π ω Σ = Σ =
Background image of page 2
3 Summary • Where X=[R,G,B] T , is a 3x 1 mean vector, is a 3x3 covariance matrix. • The method assumes that RGB color channels are independent and have the same variance . In this case = I. Where I is a 3x3 unit matrix. Learning Background Model Every new pixel is checked against all existing distributions. The match is the first distribution such that the pixel value lies within 2 standard deviations of mean. •Another way of measuring distance from a Gaussian distribution is the Mahalanobis distance i.e the match is the distribution with M-distance less than a threshold ) ( ) ( 1 j t j T j t X X d µ Σ =
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 Updating • The mean and s.d. of unmatched distributions remain unchanged. For the matched distributions they are updated as: ) ( ) ( ) 1 ( ) 1 ( , , 2 1 , , 1 , , t j t T t j t t j t j t t j t j X X X µ ρ σ + = + = • The weights are adjusted: ) ( ) 1 ( , 1 , , t j t j t j M α ω
Background image of page 4
Image of page 5
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 06/12/2011 for the course CAP 6411 taught by Professor Shah during the Spring '09 term at University of Central Florida.

Page1 / 21

Lecture-14.ppt - Change Detection, Skin Detection, Color...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online