Event Sensing on Distributed Video Sensor Networks

Similar documents
Toward Building a Robust and Intelligent Video Surveillance System: A Case Study Edward Chang and Yuan-Fang Wang

Managing and Mining Video-Sensor Data

1. Discovering Important Nodes through Graph Entropy The Case of Enron Database

Chapter 1, Introduction

Video Alignment. Literature Survey. Spring 2005 Prof. Brian Evans Multidimensional Digital Signal Processing Project The University of Texas at Austin

Title: An Integrated Design Environment to Evaluate Power/Performance Tradeoffs for Sensor Network Applications 1

Workshops. 1. SIGMM Workshop on Social Media. 2. ACM Workshop on Multimedia and Security

Multiple-Choice Questionnaire Group C

Video Alignment. Final Report. Spring 2005 Prof. Brian Evans Multidimensional Digital Signal Processing Project The University of Texas at Austin

Outline. Project Goals

An overview of infrastructures and data managers for dependable sensor networks

SENSOR SELECTION SCHEME IN TEMPERATURE WIRELESS SENSOR NETWORK

Sketch Based Image Retrieval Approach Using Gray Level Co-Occurrence Matrix

Product Characteristics Page 2. Management & Administration Page 2. Real-Time Detections & Alerts Page 4. Video Search Page 6

Action recognition in videos

Abstract. 1. Introduction

Wide Area Ontology Integration Scheme for Reasoning Agents in Surveillance Networks

Analysis of Black-Hole Attack in MANET using AODV Routing Protocol

Approach to Increase Accuracy of Multimodal Biometric System for Feature Level Fusion

SOLUTION BRIEF AN INTELLIGENT VIDEO ANALYTICS PLATFORM FOR SMART CITIES

Social Behavior Prediction Through Reality Mining

CS229: Action Recognition in Tennis

COMPARATIVE ANALYSIS OF EYE DETECTION AND TRACKING ALGORITHMS FOR SURVEILLANCE

SHRP 2 Safety Research Symposium July 27, Site-Based Video System Design and Development: Research Plans and Issues

Hybrid Face Recognition and Classification System for Real Time Environment

AN EFFECTIVE APPROACH FOR VIDEO COPY DETECTION USING SIFT FEATURES

A Comparison of SIFT and SURF

ALL ABOUT DATA AGGREGATION IN WIRELESS SENSOR NETWORKS

Gurmeet Kaur 1, Parikshit 2, Dr. Chander Kant 3 1 M.tech Scholar, Assistant Professor 2, 3

Definition, Detection, and Evaluation of Meeting Events in Airport Surveillance Videos

Human Detection. A state-of-the-art survey. Mohammad Dorgham. University of Hamburg

Energy Conservation through Sleep Scheduling in Wireless Sensor Network 1. Sneha M. Patil, Archana B. Kanwade 2

Software Design Document

A Study on Different Challenges in Facial Recognition Methods

Object Detection in Video Streams

On Demand secure routing protocol resilient to Byzantine failures

Product Characteristics Page 2. Management & Administration Page 2. Real-Time Detections & Alerts Page 4. Video Search Page 6

Link Lifetime Prediction in Mobile Ad-Hoc Network Using Curve Fitting Method

Datasheet Intrusion Detection

Recall precision graph

Cisco Kinetic for Cities Safety and Security Video Analytics (Kiwi Security)

Research Faculty Summit Systems Fueling future disruptions

USC Real-time Pattern Isolation and Recognition Over Immersive Sensor Data Streams

CV of Qixiang Ye. University of Chinese Academy of Sciences

NOVEL APPROACH TO CONTENT-BASED VIDEO INDEXING AND RETRIEVAL BY USING A MEASURE OF STRUCTURAL SIMILARITY OF FRAMES. David Asatryan, Manuk Zakaryan

DIGITAL IMAGE WATERMARKING BASED ON A RELATION BETWEEN SPATIAL AND FREQUENCY DOMAINS

Improving 3D Shape Retrieval Methods based on Bag-of Feature Approach by using Local Codebooks

Multi-Camera Occlusion and Sudden-Appearance-Change Detection Using Hidden Markovian Chains

Facial Expression Recognition using Principal Component Analysis with Singular Value Decomposition

CROSS LAYER PROTOCOL (APTEEN) USING WSN FOR REAL TIME APPLICATION

Video Privacy in the Age of Automated Surveillance

Searching Video Collections:Part I

NeoFace Watch real-time face recognition How does it work? May 2018 Global Face Recognition Centre of Excellence NEC Europe

Class 3: Advanced Moving Object Detection and Alert Detection Feb. 18, 2008

Using Geometric Blur for Point Correspondence

Clustering Methods for Video Browsing and Annotation

Event Detection through Differential Pattern Mining in Internet of Things

Feature Subset Selection Utilizing BioMechanical Characteristics for Hand Gesture Recognition

Research on outlier intrusion detection technologybased on data mining

Keywords Wavelet decomposition, SIFT, Unibiometrics, Multibiometrics, Histogram Equalization.

Chapter 9 Object Tracking an Overview

Energy Efficient Routing Using Sleep Scheduling and Clustering Approach for Wireless Sensor Network

Subject-Oriented Image Classification based on Face Detection and Recognition

Standard CIP-006-4c Cyber Security Physical Security

Ad Hoc Networks: Issues and Routing

Parallel Architecture & Programing Models for Face Recognition

Detection of Moving Object using Continuous Background Estimation Based on Probability of Pixel Intensity Occurrences

Research on Intrusion Detection Algorithm Based on Multi-Class SVM in Wireless Sensor Networks

Datasheet Intrusion Detection

Dax IVMS. Introduction. Key Features:

VIDEO OBJECT SEGMENTATION BY EXTENDED RECURSIVE-SHORTEST-SPANNING-TREE METHOD. Ertem Tuncel and Levent Onural

Shape Descriptor using Polar Plot for Shape Recognition.

Energy Efficient Tracking of Land-Based Targets Using Wireless Sensor Networks

About Me. Bio: Research Interests:

Motion Tracking and Event Understanding in Video Sequences

Critique: Efficient Iris Recognition by Characterizing Key Local Variations

An Improved Genetic Algorithm based Fault tolerance Method for distributed wireless sensor networks.

A DWT, DCT AND SVD BASED WATERMARKING TECHNIQUE TO PROTECT THE IMAGE PIRACY

Cross Reference Strategies for Cooperative Modalities

Dynamic Optimization of Generalized SQL Queries with Horizontal Aggregations Using K-Means Clustering

FUZZY FOREST LEARNING BASED ONLINE FACIAL BIOMETRIC VERIFICATION FOR PRIVACY PROTECTION

JAVA Projects. 1. Enforcing Multitenancy for Cloud Computing Environments (IEEE 2012).

Authentication and Secret Message Transmission Technique Using Discrete Fourier Transformation

Carmen Alonso Montes 23rd-27th November 2015

Criminal Identification System Using Face Detection and Recognition

Chapter 5 Ad Hoc Wireless Network. Jang Ping Sheu

Palmprint Recognition Using Transform Domain and Spatial Domain Techniques

Routing Protocols in Mobile Ad-Hoc Network

IMPACT OF LEADER SELECTION STRATEGIES ON THE PEGASIS DATA GATHERING PROTOCOL FOR WIRELESS SENSOR NETWORKS

Person Detection Using Image Covariance Descriptor

An Efficient Approach for Color Pattern Matching Using Image Mining

State-Based Synchronization Protocol in Sensor Networks

Feature-level Fusion for Effective Palmprint Authentication

Clustering in Big Data Using K-Means Algorithm

Haresh D. Chande #, Zankhana H. Shah *

THE description and representation of the shape of an object

Multi-feature face liveness detection method combining motion information

Surveillance System with Mega-Pixel Scalable Transcoder

Expanding gait identification methods from straight to curved trajectories

ANALYSIS AND EVALUATION OF DISTRIBUTED DENIAL OF SERVICE ATTACKS IDENTIFICATION METHODS

Transcription:

Event Sensing on Distributed Video Sensor Networks Edward Chang Associate Professor Department of Electrical and Computer Engineering University of California

Collaborator Prof. Yuan-Fang Wang Industry Partners Proximex ArgusSenses Intel Students Acknowledgement Ankur Jain (lead), Diana Han, Dan Koppel, Kyle Kakligian, Arun Qamra, David Trainor, Tomy Tsai

Berkeley smart dust project Stanford multi-camera array

Desired Features [refs #1,7,8,12] Independent Cameras Multiple Cameras Static, Wired Cameras Mobile, Wireless, PTZ cameras Manual Event Recognition Automatic or Semi-automatic Event Recognition Tape-based Storage RAID-based Storage No Protocol Issues Smart Semantic-based Protocols

Potential Applications and Needs Applications Emergency search and rescue in natural disaster Deterrence of cross-border illegal activities Reconnaissance and intelligence gathering in digital battlefields Needs Rapid deployment, dynamic configuration, and continuous operations Robust and real-time data fusion and analysis Intelligent event modeling and recognition

advertisement of objectvideo.com Wide-area Surveillance

Surveillance Scenarios (1) (1) Intrusion Detection (2) (2) Passenger Screening Monitor and alert on tailgating, loitering, exit/closed entry, other unauthorized access (3) (3) Perimeter Monitoring Use biometric facial recognition to identify individuals of interest through existing closed circuit TV surveillance (4) (4) Unattended Baggage Z Z Object tracking and biometric facial recognition to determine vehicles and humans exhibiting suspicious behavior Copyright 2004 Proximex Corp. Identify unattended baggage (or other objects) left for long periods of time

Intrusion Detection Loitering, Tailgating 2 Raise low severity alert!! 0 Secure entrance 1 Suspect loiters X seconds, Loitering pattern recognized 8 Correlate alerts and raise critical severity alert if if it s it s the same person! 3 Capture high resolution facial image & soft biometric data 6 Suspect tailgates authorized person in in to to secure area. System captures suspect s high resolution facial image & soft biometric data 7 Raise medium severity alert!! 4 Match watch-list 5 Raise critical severity alert if if there s a match!! Copyright 2004 Proximex Corp.

Problem Statement Video Surveillance with Multiple cameras Mobile, wireless networks Intelligent content analysis Real-time query processing Focus of Our Current Work Event sensing including Motion detection Data representation, and Event recognition Sensor network management for Bandwidth and power resource conservation Dynamic routing

Research and Development Framework Event detection Far-field coordination and update Near-field sensor data fusion Event representation Hierarchical multiple levels of detail Invariant insensitive to incidental changes Event recognition Temporally correlated event signature Imbalanced training set

System Architecture s 1 s 2 Multi-camera Data Fusion s Multi-level level Sequence Descriptor Event Recognition..... s n! Security Alert

Validation Scenario p 1 t ) = ( x ( t ), y ( t )) ( 1 1 T z 1 x 1 y 1 Slave station y m Internet x m z m P ( t ) = ( X ( t ), Y ( t ), Z ( t )) T Z Master station z 2 y 2 Y X x 2

Event Detection: Coordination and Update [Refs #5, 6] Dual Kalman filters Minimizing Bandwidth and power consumption under pre-specified accuracy constraints Update necessary only when predications diverge Cache dynamic algorithms instead of static data

Event Detection: Sensor Data Fusion [Refs #2, 9] Sensing coordination and intelligent data fusion Two-level hierarchy of Kalman filter Top level (feed backward) Fill in missing or occluded trajectory pieces Camera pose & frame rate control z x (0) (0) () t = T (0) x (0) real world p = p & p&& (0) (0) (0) X Internet Bottom level (feed forward) Summarize trajectories in local state vectors X = T Merge state vectors from multiple cameras through registration parameters Slave station world real z ( i) () t Master fusion station ( 0) x P X= P& P&& (0) X= T Slave station x ( i) p = p & p&& ( i) ( i) ( i) z ( m 1) x () t ( m 1) world real x ( m 1) = T ( m 1 ) real ( m 1) x ( p ( = p & ( p&& world ( m 1) X Slave station m 1) m 1) m 1)

Sequence Data Learning [Ref #3] Sequence Sequence Representation Numeric-valued Representation Symbolic-valued Representation Discrete Fourier Transform Wavelet Piece-wise Linear Singular Value Decomposition Natural Language Strings

Requirements Different Event Queries of Different Granularities Coarse: temporal pattern query Fine: exact match Same Events under Different Representations Circling behavior in daytime (normal) Circling behavior with abrupt turn (suspicious) Circling behavior during midnight (suspicious) Multi-level Representation Select proper resolutions in a concept and context dependent way

Kernel Design Steps 1. Design primary kernel based on symbolic description 2. Design secondary kernels based on multi-attribute vector description 3. Fuse kernels

Kernel Fusion Definition x: input vector; y: event label {K 1,,K D } be a set of D kernels for D representations K d is a (dis)similarity matrix of the training instances K d should be positive semi-definite! Kernel Fusion K fusion (x,y)= F(K 1,,K D ) F: Tensor product F: Weighted sum F: Non-linear super kernel

Event Recognition: Imbalanced Data Set [Refs #4, 10] Negative samples significantly outnumber positive samples Bayesian risk associated with false negative significantly outweighs false positive Adaptive conformal mapping at decision boundary

Sensor Network Management [Ref #8] Bit-per-joule metric Multi-hop routing Enclosure algorithm for dynamic neighbor discovery

Experimental Results: Semantic Indexing [Refs #9, 11]

Experimental Results: Statistical Learning

Concluding Remarks Multi-sensor registration and fusion Hierarchical and invariant event representation Sequence data learning using imbalanced data set Mobile camera network bandwidth and power consumption management

References 1. A collaborative Camera System for Surveillance, A. Jain, Yuan-Fang Wang, E. Y. Chang, et al., UCSB Technical Report, November 2004. 2. The Anatomy of a Multi-camera Security Surveillance System, L. Jiao, Y. Wu, G. Wu, E. Y. Chang, Y.-F. Wang, ACM Multimedia System Journal Special Issue, Volume 10, Number 2, October 2004. 3. Distance Function Design and Fusion for Sequence Data, Y. Wu and E. Y. Chang, ACM International Conference on Information and Knowledge Management (CIKM), DC, November 2004 (19% accepted). 4. Aligning Boundary in Kernel Space for Learning Imbalanced Dataset, G. Wu and E. Y. Chang, IEEE International Conference on Data Mining (ICDM), United Kingdom, November 2004 (39/452 or 9% accepted). 5. Adaptive Sampling for Sensor Networks, A. Jain and E. Y. Chang, International Workshop on Data Management for Sensor Networks in conjunction with VLDB, Toronto, August 2004. 6. Adaptive Stream Resource Management Using Kalman Filters, A. Jain, E. Y. Chang, and Y.-F. Wang, ACM International Conference on Management of Data (SIGMOD), pp.11-22, Paris, June 2004 (16% accepted). 7. Toward Building a Robust and Intelligent Video Surveillance System: A Case Study (invited paper), E. Y. Chang, Y.-F. Wang, and I-J. Wang, IEEE International Conference on Multimedia, Taipei, June 2004. 8. Distributed Video Data Fusion and Mining, E.Y. Chang, Y.-F. Wang, and Volkan Rodoplu, SPIE Conf. on Defense and Security --- Sensors, and Command, Control, Communications, and Intelligence Technologies for Homeland Security and Homeland Defense (co-chaired by DARPA, Dod/Doj), Orlando, April 2004. 9. Multi-camera Spatio-temporal Fusion and Biased Sequence-data Learning for Security Surveillance, G. Wu, Y. Wu, L. Jiao, Y.-F. Wang, and E. Y. Chang, ACM International Conference on Multimedia, pp. 528-538, Berkeley, November 2003 (17% accepted). 10. Adaptive Feature-Space Conformal Transformation for Imbalanced-Data Learning, G. Wu and E.Y. Chang, The Twentieth International Conference on Machine Learning (ICML), pp.816-823, Washington DC, August 2003. 11. Invariant Feature Extraction and Biased Statistical Inference for Video Surveillance, Y. Wu, L. Jiao, G. Wu, E. Y. Chang, and Y.-F. Wang, IEEE International Conference on Advanced Video and Signal Based Surveillance, pp.284-289, Miami, July 2003. 12. A Framework for Detecting Hazardous Events, Y. Wu, G. Wu, and E.Y. Chang, IS&T/SPIE International Conference on Storage and Retrieval for Media Databases, San Jose, January 2003. 13. More on http://www.db.stanford.edu/~echang/