Tracking in flussi video 3D. Ing. Samuele Salti



Similar documents
Practical Tour of Visual tracking. David Fleet and Allan Jepson January, 2006

Behavior Analysis in Crowded Environments. XiaogangWang Department of Electronic Engineering The Chinese University of Hong Kong June 25, 2011

Visual Tracking. Frédéric Jurie LASMEA. CNRS / Université Blaise Pascal. France. jurie@lasmea.univ-bpclermoaddressnt.fr

The Visual Internet of Things System Based on Depth Camera

VEHICLE LOCALISATION AND CLASSIFICATION IN URBAN CCTV STREAMS

Vision based Vehicle Tracking using a high angle camera

Mean-Shift Tracking with Random Sampling

Traffic Flow Monitoring in Crowded Cities

A Reliability Point and Kalman Filter-based Vehicle Tracking Technique

Classifying Manipulation Primitives from Visual Data

Automatic parameter regulation for a tracking system with an auto-critical function

Real-time Visual Tracker by Stream Processing

Video Surveillance System for Security Applications

3D Model based Object Class Detection in An Arbitrary View

VEHICLE TRACKING USING ACOUSTIC AND VIDEO SENSORS

IMPLICIT SHAPE MODELS FOR OBJECT DETECTION IN 3D POINT CLOUDS

A Comparative Study between SIFT- Particle and SURF-Particle Video Tracking Algorithms

Probabilistic Latent Semantic Analysis (plsa)

Bildverarbeitung und Mustererkennung Image Processing and Pattern Recognition

Local features and matching. Image classification & object localization

Vehicle Tracking in Occlusion and Clutter

Limitations of Human Vision. What is computer vision? What is computer vision (cont d)?

Cees Snoek. Machine. Humans. Multimedia Archives. Euvision Technologies The Netherlands. University of Amsterdam The Netherlands. Tree.

Real-Time Tracking of Pedestrians and Vehicles

Interactive Offline Tracking for Color Objects

Object tracking & Motion detection in video sequences

Tracking and integrated navigation Konrad Schindler

Component Ordering in Independent Component Analysis Based on Data Power

A Robust Multiple Object Tracking for Sport Applications 1) Thomas Mauthner, Horst Bischof

Novel Probabilistic Methods for Visual Surveillance Applications

How does Person Identity Recognition Help Multi-Person Tracking?

Object Recognition. Selim Aksoy. Bilkent University

Taking Inverse Graphics Seriously

OBJECT TRACKING USING LOG-POLAR TRANSFORMATION

Vehicle Tracking by Simultaneous Detection and Viewpoint Estimation

A Movement Tracking Management Model with Kalman Filtering Global Optimization Techniques and Mahalanobis Distance

Are we ready for Autonomous Driving? The KITTI Vision Benchmark Suite

Demo: Real-time Tracking of Round Object

Real-Time Camera Tracking Using a Particle Filter

The use of computer vision technologies to augment human monitoring of secure computing facilities

Optical Flow. Shenlong Wang CSC2541 Course Presentation Feb 2, 2016

Who are you? Learning person specific classifiers from video

The Scientific Data Mining Process

EE 570: Location and Navigation

Segmentation & Clustering

Visual Vehicle Tracking Using An Improved EKF*

FACE RECOGNITION BASED ATTENDANCE MARKING SYSTEM

Deterministic Sampling-based Switching Kalman Filtering for Vehicle Tracking

An Experimental Comparison of Online Object Tracking Algorithms

Professor, D.Sc. (Tech.) Eugene Kovshov MSTU «STANKIN», Moscow, Russia

Learning Motion Categories using both Semantic and Structural Information

To determine vertical angular frequency, we need to express vertical viewing angle in terms of and. 2tan. (degree). (1 pt)

Real Time Target Tracking with Pan Tilt Zoom Camera

Master s thesis tutorial: part III

Automatic Maritime Surveillance with Visual Target Detection

Understanding and Applying Kalman Filtering

Subspace Analysis and Optimization for AAM Based Face Alignment

PEST - Beyond Basic Model Calibration. Presented by Jon Traum

QUALITY ENGINEERING PROGRAM

Tracking and Recognition in Sports Videos

EFFICIENT VEHICLE TRACKING AND CLASSIFICATION FOR AN AUTOMATED TRAFFIC SURVEILLANCE SYSTEM

False alarm in outdoor environments

MetropoGIS: A City Modeling System DI Dr. Konrad KARNER, DI Andreas KLAUS, DI Joachim BAUER, DI Christopher ZACH

Recognizing Cats and Dogs with Shape and Appearance based Models. Group Member: Chu Wang, Landu Jiang

3D Vehicle Extraction and Tracking from Multiple Viewpoints for Traffic Monitoring by using Probability Fusion Map

8. Linear least-squares

Statistical Models in Data Mining

Incremental PCA: An Alternative Approach for Novelty Detection

Rafael Martín & José M. Martínez

Localization of Mobile Robots Using Odometry and an External Vision Sensor

Modelling, Extraction and Description of Intrinsic Cues of High Resolution Satellite Images: Independent Component Analysis based approaches

Object Tracking for Laparoscopic Surgery Using the Adaptive Mean-Shift Kalman Algorithm

FLEXSYS Motion-based Traffic Analysis and Incident Detection

A Study on SURF Algorithm and Real-Time Tracking Objects Using Optical Flow

Speed Performance Improvement of Vehicle Blob Tracking System

Probability and Random Variables. Generation of random variables (r.v.)

MusicGuide: Album Reviews on the Go Serdar Sali

Colorado School of Mines Computer Vision Professor William Hoff

Feature Tracking and Optical Flow

ARC 3D Webservice How to transform your images into 3D models. Maarten Vergauwen

Christfried Webers. Canberra February June 2015

Human behavior analysis from videos using optical flow

View-Invariant Dynamic Texture Recognition using a Bag of Dynamical Systems

Introduction to Mobile Robotics Bayes Filter Particle Filter and Monte Carlo Localization

A Learning Based Method for Super-Resolution of Low Resolution Images

STA 4273H: Statistical Machine Learning

Tracking based on graph of pairs of plots

Tracking And Object Classification For Automated Surveillance

Transcription:

Seminari XXIII ciclo Tracking in flussi video 3D Ing. Tutors: Prof. Tullio Salmon Cinotti Prof. Luigi Di Stefano

The Tracking problem Detection Object model, Track initiation, Track termination, Tracking Object motion model, Model update, Multi-target tracking / Data association Occlusion handling, Combinatorial problem (Exponential complexity with growing number of targets),

2D Tracking State of the art performances in 2D videos Main idea: Tracking-by- Detection Reliable detector used in every frame: Implicit Shape Model (ISM), Histogram-of-Gradient (HOG), etc Tracking reformulated as data association across frames Limitations People pose Occlusions & clutter Illumination changes Output 2D Liebe & al., IJCV 08, Breitentesin & al., ICCV 09

Why not just one image? Analyzing a single view is not possible to unambiguously reconstruct the 3D structure of the scene This is due to effects of the perspective projection that maps points of a 3D space in a 2D space (the image plane of the camera)

3D acquisition devices

3D data and previous work Most exploited approach Camera calibrated wrt the ground plane People detected with background subtraction 2D projection of 3D data Tracking in 2D plan view Limitations Assume static camera Requires a background model Requires calibration Bottom-up approach Beymer & Konolige 2000 Iocchi & Bolles ICIP 2005 Harville & Li, CVPR 04 Yous & al., ECCV WS 2008

My contribution Design an enhanced people detector, exploiting the full potential of 3D data Toward this goal propose a new 3D descriptor of local shape suitable for our task Design a theoretically sound and adaptive way to merge 2D and 3D info for the purpose of people detection (i.e. object category recognition) Plug this in a tracking framework conceived for time critical, online applications No global optimization More emphasis on tracking than on data association Recursive Bayesian Estimation (RBE) methods Enhance RBE via machine learning

3D shape descriptor Our proposal dubbed HON: Histogram of Normals Designed to be Fast Robust to noise and clutter Robust to sampling density variations Definition of a new, robust way to compute an invariant local reference frame Inspired to successful approaches for 2D texture description Lowe, IJCV 04 cos θ

HON: Results on noise and clutter recall 1-precision

HON: Results on sampling density recall 1-precision

My contribution Design an enhanced people detector, exploiting the full potential of 3D data Toward this goal propose a new 3D descriptor of local shape suitable for our task Design a theoretically sound and adaptive way to merge 2D and 3D info for the purpose of people detection (i.e. object category recognition) Plug this in a tracking framework conceived for time critical, online applications No global optimization More emphasis on tracking than on data association Recursive Bayesian Estimation (RBE) methods Enhance RBE via machine learning

Recursive Bayesian Estimation RBE provides a theoretically sound conceptual solution to the problem of state estimation in presence of uncertainty. RBE is widely employed in the context of Visual Tracking and Motion Analysis. In this framework the system is completely specified by a first order Markov model compound of a transition model in state space x (, k = fk xk 1 υ ) p x k k xk a measurement model zk = hk ( xk, ηk) p zk xk an initial state x 0 p ( x0 ) Practical instantiations ( 1 ) ( ) the Kalman filter (Linear & Gaussian scenario, optimal solution) the particle filter (Non-Linear / Non-Gaussian scenario, sub-optimal solution)

Motivations A major limitation of RBE is the requirement to a priori specify the transition model. In most cases this model is unknown and is empirically selected among a restricted set of standard ones or it is learned off-line Both approaches do not allow for changing the transition model trough time, although this would be beneficial and neither the conceptual solution nor the solving algorithms require this.

Proposal In case of a completely observable system, we propose to learn the transition model on-line. In such a case, the transition model is directly related to the dynamics exhibited by the measures. Hence, it is possible to exploit their temporal evolution in order to learn the function f x υ, and, implicitly, the PDF x x. ( ), k z1: k 1 k 1 k p k z Furthermore, we propose to learn the motion model using Support Vector Machine in ε-regression mode (SVR) SVR theoretical properties minimize the risk of overfitting SVR can learn non-linear mapping effectively via the kernel trick SVR can be trained very efficiently exploiting SMO ( ) 1: k 1 k k 1

Support Vector Kalman RBE in the linear & Gaussian scenario becomes: In this case, the PDF we want to estimate becomes ( ) = ( ; μ; Σ ) = ( ; ; ) p x x N x N x Fx Q k z 1: k 1 k k 1 k k k k 1 k Therefore, we use SVRs to estimate the transition matrix F k the associated noise covariance matrix, Q k

Simulations

Mean Shift Tracking

Future work Design an enhanced people detector, exploiting the full potential of 3D data Toward this goal propose a new 3D descriptor of local shape suitable for our task Design a theoretically sound and adaptive way to merge 2D and 3D info for the purpose of people detection (i.e. object category recognition) Plug this in a tracking framework conceived for time critical, online applications No global optimization More emphasis on tracking than on data association Recursive Bayesian Estimation (RBE) methods Enhance RBE via machine learning