Inferring Temporal Order of Images from 3D Structure

Slides:



Advertisements
Similar presentations
Introduction to Algorithms Graph Algorithms
Advertisements

A Robust Super Resolution Method for Images of 3D Scenes Pablo L. Sala Department of Computer Science University of Toronto.
Top-Down & Bottom-Up Segmentation
Constraint Satisfaction Problems
Some Graph Algorithms.
Guiding Semi- Supervision with Constraint-Driven Learning Ming-Wei Chang,Lev Ratinow, Dan Roth.
For Internal Use Only. © CT T IN EM. All rights reserved. 3D Reconstruction Using Aerial Images A Dense Structure from Motion pipeline Ramakrishna Vedantam.
Review: Constraint Satisfaction Problems How is a CSP defined? How do we solve CSPs?
NUS CS5247 Motion Planning for Camera Movements in Virtual Environments By Dennis Nieuwenhuisen and Mark H. Overmars In Proc. IEEE Int. Conf. on Robotics.
Onur G. Guleryuz & Ulas C.Kozat DoCoMo USA Labs, San Jose, CA 95110
Patch to the Future: Unsupervised Visual Prediction
1.Introduction 2.Article [1] Real Time Motion Capture Using a Single TOF Camera (2010) 3.Article [2] Real Time Human Pose Recognition In Parts Using a.
Qualifying Exam: Contour Grouping Vida Movahedi Supervisor: James Elder Supervisory Committee: Minas Spetsakis, Jeff Edmonds York University Summer 2009.
1 Modularity and Community Structure in Networks* Final project *Based on a paper by M.E.J Newman in PNAS 2006.
Discrete-Continuous Optimization for Large-scale Structure from Motion David Crandall, Andrew Owens, Noah Snavely, Dan Huttenlocher Presented by: Rahul.
Graph Traversals Reading Material: Chapter 9. Graph Traversals Some applications require visiting every vertex in the graph exactly once. The application.
1 Structure from Motion with Unknown Correspondence (aka: How to combine EM, MCMC, Nonlinear LS!) Frank Dellaert, Steve Seitz, Chuck Thorpe, and Sebastian.
Anagh Lal Monday, April 14, Chapter 9 – Tree Decomposition Methods Anagh Lal CSCE Advanced Constraint Processing.
Segmentation CSE P 576 Larry Zitnick Many slides courtesy of Steve Seitz.
Lecture 11: Structure from motion, part 2 CS6670: Computer Vision Noah Snavely.
Lecture 16 CSE 331 Oct 9, Announcements Hand in your HW4 Solutions to HW4 next week Remember next week I will not be here so.
Large-Scale Deduplication with Constraints using Dedupalog Arvind Arasu et al.
3D reconstruction of cameras and structure x i = PX i x’ i = P’X i.
UNC Chapel Hill M. C. Lin Overview of Last Lecture About Final Course Project –presentation, demo, write-up More geometric data structures –Binary Space.
CS 223B Assignment 1 Help Session Dan Maynes-Aminzade.
Kyle Heath, Natasha Gelfand, Maks Ovsjanikov, Mridul Aanjaneya, Leo Guibas Image Webs Computing and Exploiting Connectivity in Image Collections.
Brief overview of ideas In this introductory lecture I will show short explanations of basic image processing methods In next lectures we will go into.
3-D Scene u u’u’ Study the mathematical relations between corresponding image points. “Corresponding” means originated from the same 3D point. Objective.
Image Segmentation Rob Atlas Nick Bridle Evan Radkoff.
3D Fingertip and Palm Tracking in Depth Image Sequences
CSSE463: Image Recognition Day 34 This week This week Today: Today: Graph-theoretic approach to segmentation Graph-theoretic approach to segmentation Tuesday:
Nattee Niparnan. Graph  A pair G = (V,E)  V = set of vertices (node)  E = set of edges (pairs of vertices)  V = (1,2,3,4,5,6,7)  E = ((1,2),(2,3),(3,5),(1,4),(4,
Tzu ming Su Advisor : S.J.Wang MOTION DETAIL PRESERVING OPTICAL FLOW ESTIMATION 2013/1/28 L. Xu, J. Jia, and Y. Matsushita. Motion detail preserving optical.
Chapter 14: SEGMENTATION BY CLUSTERING 1. 2 Outline Introduction Human Vision & Gestalt Properties Applications – Background Subtraction – Shot Boundary.
Image-based Plant Modeling Zeng Lanling Mar 19, 2008.
BUILDING EXTRACTION AND POPULATION MAPPING USING HIGH RESOLUTION IMAGES Serkan Ural, Ejaz Hussain, Jie Shan, Associate Professor Presented at the Indiana.
Object Stereo- Joint Stereo Matching and Object Segmentation Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on Michael Bleyer Vienna.
CSP Examples Tamara Berg CS 560 Artificial Intelligence Many slides throughout the course adapted from Svetlana Lazebnik, Dan Klein, Stuart Russell, Andrew.
Discussion #32 1/13 Discussion #32 Properties and Applications of Depth-First Search Trees.
Tracking People by Learning Their Appearance Deva Ramanan David A. Forsuth Andrew Zisserman.
Computer Vision, Robert Pless
Learning Spectral Clustering, With Application to Speech Separation F. R. Bach and M. I. Jordan, JMLR 2006.
Final Review Course web page: vision.cis.udel.edu/~cv May 21, 2003  Lecture 37.
Yizhou Yu Texture-Mapping Real Scenes from Photographs Yizhou Yu Computer Science Division University of California at Berkeley Yizhou Yu Computer Science.
Properties and Applications of Depth-First Search Trees and Forests
Using decision trees to build an a framework for multivariate time- series classification 1 Present By Xiayi Kuang.
Discriminative Training and Machine Learning Approaches Machine Learning Lab, Dept. of CSIE, NCKU Chih-Pin Liao.
Mestrado em Ciência de Computadores Mestrado Integrado em Engenharia de Redes e Sistemas Informáticos VC 15/16 – TP10 Advanced Segmentation Miguel Tavares.
ISAM2: Incremental Smoothing and Mapping Using the Bayes Tree Michael Kaess, Hordur Johannsson, Richard Roberts, Viorela Ila, John Leonard, and Frank Dellaert.
Photoconsistency constraint C2 q C1 p l = 2 l = 3 Depth labels If this 3D point is visible in both cameras, pixels p and q should have similar intensities.
Cell Segmentation in Microscopy Imagery Using a Bag of Local Bayesian Classifiers Zhaozheng Yin RI/CMU, Fall 2009.
Lecture 22: Structure from motion CS6670: Computer Vision Noah Snavely.
Scene Parsing with Object Instances and Occlusion Ordering JOSEPH TIGHE, MARC NIETHAMMER, SVETLANA LAZEBNIK 2014 IEEE CONFERENCE ON COMPUTER VISION AND.
Breadth-First Search (BFS)
A Plane-Based Approach to Mondrian Stereo Matching
Miguel Tavares Coimbra
CSSE463: Image Recognition Day 34
Using aerial images for urban planning
EECS 274 Computer Vision Stereopsis.
Structure from motion Input: Output: (Tomasi and Kanade)
Level Set Tree Feature Detection
Scale-Space Representation of 3D Models and Topological Matching
Image processing and computer vision
Image Segmentation.
A Novel Smoke Detection Method Using Support Vector Machine
CSSE463: Image Recognition Day 34
Structure from motion Input: Output: (Tomasi and Kanade)
“Traditional” image segmentation
Graph Traversals Some applications require visiting every vertex in the graph exactly once. The application may require that vertices be visited in some.
Presentation transcript:

Inferring Temporal Order of Images from 3D Structure Grant Schindler Gatech Frank Dellaert Gatech Sing Bing Kang MSR, Redmond

Outline Problem Definition Algorithm Overview Applications Things to think about

What can be done with n images?

What can be done with n images? Feature Extraction Correspondence Structure from Motion What Now?

Temporal Ordering and 4D Walkthrough 1920 1951 1966 2003

Outline Problem Definition Algorithm Overview Applications Things to think about

SFM tells us: F1 F2 C1 I1 F3 I2 C2 Camera Matrices 3D points for features Visibility of 3D points in images

C2 F3 C1 I1 I2 F1 F2 SFM info Image 1 (I1) Image 2 (I2) F1 Visible ??? F2 F3

C2 F3 C1 I1 I2 F1 F2 SFM info Image 1 (I1) Image 2 (I2) F1 Visible ??? F2 Occluded F3 Out of View

Notion of missing at that time C2 F3 C1 I1 I2 F1 F2 Notion of missing at that time SFM info Image 1 (I1) Image 2 (I2) F1 Visible ??? F2 Occluded F3 Out of View

Classification of 3D point for an Image Visible – SFM tells us Out of View – Camera Matrix tells us Missing / Occluded - ??? for an occluded point, there must be an occluder

Point ‘F’ Missing / Occluded ? Find out occluders 3D Triangulation of all visible points No occluder should occlude a visible point Visibility check for F occluders F1 F2 Camera centre occluded missing

Visibility Matrix Sij € {visible, occluded, missing, out of view } I1 ... In F1 S11 S12 … S1n F2 S21 S22 S2n Fm Sm1 Sm2 Smn Sij € {visible, occluded, missing, out of view }

Constraints of Visibility Matrix

Combinatorial Algorithm to find Best Configuration Local search method Starts at a random configuration Small moves which reduce the number of constraints violated

Issues leading to Finding Approximate Solution Problems in feature detection Mislabeling of points Triangulation strategy Inaccuracy in SFM Features occluded by undetected occluders (fog, trees etc)

Structural Segmentation from Temporal Ordering Clustering temporally coherent features Separate triangulation of each cluster Texture by projecting on images

Algorithm Overview

Possible Applications Historic Preservation Virtual Tourism Urban Planning Spatio-Temporal models as a new way to interact with a vast collection of imagery

Things to Think about Feature extraction (done manually here) Better methods for finding occluders – problems with triangulation method Very coarse structure Can have triangles for no occluders Using Goesele’s work (ICCV 2007) for structural segmentation High number of images required (this paper used 20-30 images) Validation Correspondence between the best ground truth solution and best approximate solution of ordering Increasing the scale technically and physically

An Interesting Insight…. Assume no building can be demolished once it’s built Assume every image is a node of graph Edge from A to B if A precedes B (B has visible features missing in A ) Directed Graph (acyclic in ideal case)

B1 B2 B3 A B C Input Images C A B Directed Graph (Acyclic in ideal case)

B1 B2 B3 A B C Input Images C A B Directed Graph (Acyclic in ideal case) B C A Topological Sort Solution !

More insights about Graph Model Every edge has a confidence value based on quality of features and SFM procedure In general, there can be back edges in this graph Problem to find the best topological sort maximizing the confidence measure

Graph Complexity Increases with more constraints Modeling constraints involving more than 2 images at a time - how??