Unsual Behavior Analysis and Its Application to Surveillance Syste

msYung-Tai Hsu(許詠泰 )

Jun-Wei Hsieh(謝君偉 )

Hong-Yuan Mark Liao(廖弘源 )

Introduction

• Deformable Triangulations

• Skeleton-based Posture Recognition

• Posture Recognition Using the Centroid Context

• Experiment Results

Deformable Triangulations

• P is a posture extracted in binary form by image subtraction(fig.1)

• B is the set of boundary points along the contour of P(fig.2)

• Extract some high curvature points from B(fig.3)• α(p) is the angle of a point p in B. It can be determined b

y two specified points p+ and p-.(fig.4)

fig.1 fig.2 fig.3 fig.4

Deformable Triangulations

Dmin = |B| / 30, Dmax = |B| / 20

• If α is larger than a threshold Tα (here we set it at 150), p is selected as a control point.

• If two candidates, p1 and p2 are close to each other, i.e., ||p1 – p2||<dmin, the candidate with smaller α angle is chosen as a control point.

Deformable Triangulations

ViVj

Vk

VaVb

Triangulation-based Skeleton Extraction

• P is decomposed into a set of triangle meshes Ωp

• Ωp={Ti}i=0,1,2,…,NTP-1

• Each triangle mesh Ti in Ωp has a centroid CTi

• H is defined as the head of P and it is the highest node among all the nodes.

• All the leaf nodes Li correspond to different limbs of P

• The branching nodes Bi are the key points used to decompose P into different body parts, such as the hands, feet, or torso.

Posture Recognition Using a Skeleton

• Assume SP and SD are two skeletal images extracted from a posture P and D.

• Assume DTSP is the distance map of SP.

• The value of a pixel r in DTSP is its shortest distance to all foreground pi

xels in SP.• d(r, q) is the Euclidian distance between r and q.

• |DTSP| represents the image size of DTSP

.

• SP and SD must be normalized to a unit size and their centers must be set to the origins of DTSP

and DTSD.

Posture Recognition Using a Skeleton

a) Shows the original posture.

b) It is the result of skeleton extraction.

c) Shows the resultant distance map based on (b)

Centroid Context-based Description of Postures

• Assume all postures are normalized to a unit size.• We project a sample onto a log-polar coordinate and label each mes

h. • Use m to represent the number of shells used to quantize the radial

axis and • use n to represent the number of sectors that we would like to quant

ize each shell.• The total number of bins used to construct the centroid context is m

×n.• For each centroid r of a triangle mesh of a posture, we construct a v

ector histogram hr.

• hr(k) is the number of triangle mesh centroids in the kth bin by considering r as the origin

• bink is the kth bin of the log-polar coordinate.

Centroid Context-based Description of Postures

• Given two histograms hri(k) and hrj

(k), the distance between them can be measured by a normalized intersection:

Centroid Context-based Description of Postures

• |VP| is the number of elemetns in VP.

Centroid Context-based Description of Postures

• Give two postures P and Q, the distance between their centroid contexts is measured by:

• Where w and w are the area ratios of the ith and jth body parts residing in P and Q.

Centroid Context-based Description of Postures

Posture Recognition Using the Skeleton and the Centroid Context

• Ti is the ith normal behavior with the training threshold.• q is the query posture.• ri,j is the jth key posture of the ith normal behavior with le

ngth N.

Experiment Results