Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 2014-10-07 Yeong-Jun Cho Computer Vision and Pattern Recognition,2013.

Slides:



Advertisements
Similar presentations
Recognising Panoramas M. Brown and D. Lowe, University of British Columbia.
Advertisements

3D Model Matching with Viewpoint-Invariant Patches(VIP) Reporter :鄒嘉恆 Date : 10/06/2009.
Location Recognition Given: A query image A database of images with known locations Two types of approaches: Direct matching: directly match image features.
Simultaneous surveillance camera calibration and foot-head homology estimation from human detection 1 Author : Micusic & Pajdla Presenter : Shiu, Jia-Hau.
Presented by Xinyu Chang
Hilal Tayara ADVANCED INTELLIGENT ROBOTICS 1 Depth Camera Based Indoor Mobile Robot Localization and Navigation.
Query Specific Fusion for Image Retrieval
Object Recognition using Invariant Local Features Applications l Mobile robots, driver assistance l Cell phone location or object recognition l Panoramas,
Special Topic on Image Retrieval Local Feature Matching Verification.
Image alignment Image from
Computer Vision Group, University of BonnVision Laboratory, Stanford University Abstract This paper empirically compares nine image dissimilarity measures.
Title of Presentation Author 1, Author 2, Author 3, Author 4 Abstract Introduction This is my abstract. This is my abstract. This is my abstract. This.
Object Recognition with Invariant Features n Definition: Identify objects or scenes and determine their pose and model parameters n Applications l Industrial.
Effective Image Database Search via Dimensionality Reduction Anders Bjorholm Dahl and Henrik Aanæs IEEE Computer Society Conference on Computer Vision.
Robust and large-scale alignment Image from
A Novel 2D-to-3D Conversion System Using Edge Information IEEE Transactions on Consumer Electronics 2010 Chao-Chung Cheng Chung-Te li Liang-Gee Chen.
Object retrieval with large vocabularies and fast spatial matching
Object Recognition with Invariant Features n Definition: Identify objects or scenes and determine their pose and model parameters n Applications l Industrial.
Recognising Panoramas
Geometric Optimization Problems in Computer Vision.
Linear Solution to Scale and Rotation Invariant Object Matching Professor: 王聖智 教授 Student : 周 節.
Automatic Image Alignment (feature-based) : Computational Photography Alexei Efros, CMU, Fall 2006 with a lot of slides stolen from Steve Seitz and.
Multiple Object Class Detection with a Generative Model K. Mikolajczyk, B. Leibe and B. Schiele Carolina Galleguillos.
Computing transformations Prof. Noah Snavely CS1114
Feature Matching and RANSAC : Computational Photography Alexei Efros, CMU, Fall 2005 with a lot of slides stolen from Steve Seitz and Rick Szeliski.
Keypoint-based Recognition and Object Search
כמה מהתעשייה? מבנה הקורס השתנה Computer vision.
Yuping Lin and Gérard Medioni.  Introduction  Method  Register UAV streams to a global reference image ▪ Consecutive UAV image registration ▪ UAV to.
CSE 185 Introduction to Computer Vision
Chapter 6 Feature-based alignment Advanced Computer Vision.
Keypoint-based Recognition Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem 03/04/10.
Efficient Algorithms for Robust Feature Matching Mount, Netanyahu and Le Moigne November 7, 2000 Presented by Doe-Wan Kim.
Fast Approximate Energy Minimization via Graph Cuts
Advanced Computer Vision Feature-based Alignment Lecturer: Lu Yi & Prof. Fuh CSIE NTU.
Characterizing activity in video shots based on salient points Nicolas Moënne-Loccoz Viper group Computer vision & multimedia laboratory University of.
A Local Adaptive Approach for Dense Stereo Matching in Architectural Scene Reconstruction C. Stentoumis 1, L. Grammatikopoulos 2, I. Kalisperakis 2, E.
A Statistical Approach to Speed Up Ranking/Re-Ranking Hong-Ming Chen Advisor: Professor Shih-Fu Chang.
Feature-Based Stereo Matching Using Graph Cuts Gorkem Saygili, Laurens van der Maaten, Emile A. Hendriks ASCI Conference 2011.
A Region Based Stereo Matching Algorithm Using Cooperative Optimization Zeng-Fu Wang, Zhi-Gang Zheng University of Science and Technology of China Computer.
Computer Vision : CISC 4/689 Going Back a little Cameras.ppt.
18 th August 2006 International Conference on Pattern Recognition 2006 Epipolar Geometry from Two Correspondences Michal Perďoch, Jiří Matas, Ondřej Chum.
Source: Computer Vision and Pattern Recognition Workshops (CVPRW), 2010 IEEE Computer Society Conference on Author: Paucher, R.; Turk, M.; Adviser: Chia-Nian.
CSE 185 Introduction to Computer Vision Feature Matching.
Bundling Features for Large Scale Partial-Duplicate Web Image Search Zhong Wu ∗, Qifa Ke, Michael Isard, and Jian Sun Microsoft Research.
Title Authors Introduction Text, text, text, text, text, text Background Information Text, text, text, text, text, text Observations Text, text, text,
776 Computer Vision Jan-Michael Frahm Spring 2012.
2D to 3D Conversion Using 3D Database For Football Scenes Kiana Calagari Final Project of CMPT880 July 2013.
A Discriminatively Trained, Multiscale, Deformable Part Model Yeong-Jun Cho Computer Vision and Pattern Recognition,2008.
COS 429 PS3: Stitching a Panorama Due November 10 th.
CSCI 631 – Foundations of Computer Vision March 15, 2016 Ashwini Imran Image Stitching.
DETECTION OF COPY MOVE FORGERY IN DIGITAL IMAGES.
EE 7730 Parametric Motion Estimation. Bahadir K. Gunturk2 Parametric (Global) Motion Affine Flow.
CSCI 631 – Foundations of Computer Vision March 15, 2016 Ashwini Imran Image Stitching Link: singhashwini.mesinghashwini.me.
776 Computer Vision Jan-Michael Frahm Spring 2012.
SIFT Scale-Invariant Feature Transform David Lowe
Summary of “Efficient Deep Learning for Stereo Matching”
Capturing, Processing and Experiencing Indian Monuments
Content-based Image Retrieval
Efficient Sampling Methods for Robust Estimation Problems
Homography From Wikipedia In the field of computer vision, any
Christian Wolf 1, Jean-Michel Jolion 2, Walter G
TITLE Authors Institution RESULTS INTRODUCTION CONCLUSION AIMS METHODS
Cheng-Ming Huang, Wen-Hung Liao Department of Computer Science
Object recognition Prof. Graeme Bailey
Rob Fergus Computer Vision
Feature Matching and RANSAC
Large Scale Image Deduplication
SIFT.
CSE 185 Introduction to Computer Vision
Recognition and Matching based on local invariant features
Presentation transcript:

Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition Yeong-Jun Cho Computer Vision and Pattern Recognition,2013

 Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition –Introduction –Methods –Results –Conclusion  Conclusion Contents 2

Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition CVPR

 Introduction Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 4 2D-to-3D matching 을 통한 3D Object recognition

 Introduction Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 5 Query Image 와 3D object models 과의 모든 correspondences 를 구함 2D feature 로는 DAISY 사용 / Searching 기법으로는 ANN(approximate nearest neighbor) 사용 Inlier Outlier Correspondences

 Introduction Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 6 모델별로 RANSAC 을 통해 최종 inlier 를 선별 최초 matching 결과의 outlier 가 많으면 많은 RANSAC iteration 을 요구함. ( 수행 시간 증가 ) 뿐만 아니라, RANSAC 정확도가 떨어져 recognition recall 이 떨어질 수 있음. ( 인식 정확도 하락 ) Inlier Outlier Correspondences

 Introduction Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 7 따라서, 최초 matching 시의 Outlier 를 빠르고 효과적으로 제거하는 기법을 제안 (Correspondence filtering) ▶ 수행속도 개선, 인식 정확도 개선 Inlier Outlier

 Introduction – 문제 정의 (outlier 종류 기술 ) Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 8 배경과 matching 다른 model 과 matching 같은 model 과 matching 되었으나, 올바르지 않은 위치에 matching

 Introduction – 문제 정의 (outlier 종류 기술 ) Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 9 배경과 matching 다른 model 과 matching 같은 model 과 matching 되었으나, 올바르지 않은 위치에 matching Statistics & geometric cues 를 통한 outlier 제거 Global filtering Local filtering

 Methods –Local filtering for removing Authors observed thatare irregularly distributed Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 10

 Methods –Local filtering for removing Authors observed thatare irregularly distributed Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 11

 Methods –Local filtering for removing Authors observed thatare irregularly distributed 2D local consistency check Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 12 scene

 Methods –Local filtering for removing Authors observed thatare irregularly distributed 2D local consistency check Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 13 scene

 Methods –Local filtering for removing Authors observed thatare irregularly distributed 2D-3D local consistency check Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 14 2D-3D local consistency check

 Methods –Global filtering for removing Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 15

 Methods Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 16

 Methods Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 17

 Methods Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 18

 Methods Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 19 상대적으로 강한 연결이 되지 않은 vertex 는 로 판단하여 제거

 Experimental results Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 20

 Experimental results Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 21

 Conclusion – 각 correspondence 분포를 고려한 Local filtering 과 –Pairwise 한 위치 관계를 고려한 Global filtering 을 통한 outlier correspondences 제거 > 수행 속도 향상 및 인식 정확도 향상 Efficient 2D-to-3D Correspondence Filtering for Scalable 3D Object Recognition 22

Q & A 23