Lecture-10

Lecture-10 - Alper Yilmaz Motivation Change Detection, Skin...

Info iconThis preview shows pages 1–4. Sign up to view the full content.

View Full Document Right Arrow Icon
Alper Yilmaz UCF Computer Vision Lab. 1 Change Detection, Skin Detection Lecture-10 UCF Computer Vision Lab. 2 Motivation ! Detection of interesting objects in videos is the first step in the process of automated surveillance. • Focus of attention method greatly reduces the processing time required for tracking and activity recognition. UCF Computer Vision Lab. 3 Introduction ! Objectives: ! Given a sequence of images from a stationary camera identify pixels comprising ‘interesting’ objects. • General Solution –Model properties of the scene (e.g. color, texture e.t.c) at each pixel. – All independently moving objects are interesting! –Significant change in the properties indicates an interesting change. UCF Computer Vision Lab. 4 Introduction ! Problems in Realistic situations: ! Moving but uninteresting objects ! e.g. trees, flags or grass. – Long term illumination changes •e.g. time of day. – Quick illumination changes •e.g. cloudy weather – Shadows – Other Physical changes in the background •Dropping or picking up of objects – Initialization UCF Computer Vision Lab. 5 Segmenting Background UCF Computer Vision Lab. 6 Difference Pictures ! Jain, R. and Nagel, H. 1979. ``On the analysis of accumulative difference pictures from image sequences of real world scenes”. IEEE Trans. on Pattern Analysis and Machine Intelligence 1, 2, pp 206-214.
Background image of page 1

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

View Full DocumentRight Arrow Icon
Alper Yilmaz UCF Computer Vision Lab. 2 UCF Computer Vision Lab. 7 Background Subtraction - = ! Problem: Choosing a threshold ! Pixel is foreground if I 1 ( x , y )- 2 ( , ) ! ! otherwise background? ! What is the correct value of ! ? UCF Computer Vision Lab. 8 Setting a Threshold ! = 10 ! = 20 ! = 50 ! = 100 ! = 200 ! = 300 UCF Computer Vision Lab. 9 MODELING PIXEL INTENSITIES WITH A NORMAL DISTRIBUTION Each pixel intensity can be modeled by a Normal Distribution, defined in terms of a mean " and variance # 2 , as as N ( " , # 2 ). ). " and and # are called parameters. are called parameters. Useful when you wish to establish membership of a pixel to one of several models. N ( " , # 2 ) is a probability distribution ) is a probability distribution function defined by: ) 2 ( ) ( 2 2 2 2 2 1 ) , | ( # " $# % % & x e f x UCF Computer Vision Lab. 10 Bi-variate Normal Distribution ! If we were interested in r-g , or g-b , or r- b ! The mean can be updated over time simply as UCF Computer Vision Lab. 11 Covariance ( ) * * + , & - 0 0 ( ) * * + , & - 2 1 0 0 ( ) * * + , & - 22 21 12 11 UCF Computer Vision Lab. 12 Tri-variate Normal
Background image of page 2
Alper Yilmaz UCF Computer Vision Lab. 3 UCF Computer Vision Lab. 13 Method I: P finder !
Background image of page 3

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

View Full DocumentRight Arrow Icon
Image of page 4
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 9

Lecture-10 - Alper Yilmaz Motivation Change Detection, Skin...

This preview shows document pages 1 - 4. Sign up to view the full document.

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