On Clustering Images of Objects

Size: px
Start display at page:

Download "On Clustering Images of Objects"

Transcription

1 On Clustering Images of Objects Jongwoo Lim University of Illinois at Urbana-Champaign Honda Research Institute

2 Personal Introduction Education B.S. Seoul National University, 1993~1997 Ph.D. student at University of Illinois at Urbana-Champaign, 2000 ~ M.S. University of Illinois at Urbana-Champaign, 2003 Visiting, University of California at San Diego, 2002~2005 Experience Honda Research Institute, 2003, 2005~ Summer Intern : ASIMO face detection/recognition/tracking system Senior Research Scientist UIUC Active Space project, 2001~2005 Research Assistant : people tracking system Triton Tech, 1997~2000 Assistant Manager / Programmer : electronic document management system [2]

3 Publications Beyond Pairwise Clustering S. Agarwal, J. Lim, L. Zelnik, P. Perona, D. Kriegman, S. Belongie / CVPR 2005, vol. 2, pp CVPR05 Image Clustering with Metric, Local Linear Structure and Affinity Symmetry J. Lim, J. Ho, M-H. Yang, K-C. Lee, D. Kriegman / ECCV 2004, vol. 1, pp Clustering Appearances of Objects Under Varying Illumination Conditions J. Ho, M-H. Yang, J. Lim, K-C. Lee, D. Kriegman / CVPR 2003, vol 1, pp PhD Thesis Passive Photometric Stereo from Motion J. Lim, J. Ho, M-H. Yang, D. Kriegman / ICCV 2005, to appear ICCV05, accepted A Direct Method for Modeling Non-rigid Motion with Thin Plate Spline J. Lim, M-H. Yang / CVPR 2005, vol. 1, pp CVPR05 Incremental Learning for Visual Tracking J. Lim, D. Ross, R-S. Lin, M-H. Yang / NIPS 2004 Adaptive Discriminative Generative Model and Its Applications R-S. Lin, D. Ross, J. Lim, M-H. Yang / NIPS 2004 Probabilistic Visual Tracking with Incremental Subspace Update D. Ross, J. Lim, M-H. Yang / ECCV 2004, vol. 2, pp Tracking Humans Using Prior and Learned Representations of Shape and Appearance J. Lim, D. Kriegman / FGR 2004, pp Master Thesis [3]

4 Outline Introduction Generic Clustering Algorithms Normalized Cut / Spectral Clustering Hypergraph Clustering Clustering Images When illumination condition changes. When pose (viewing direction) changes. When pose and illumination changes simultaneously. Conclusion [4]

5 What is the Clustering Problem? Grouping data elements with similar characteristics. E.g., Perceptual grouping of points in 2D Position or distance in the space. [Zelnik,Perona NIPS04] [5]

6 Another Clustering Example Clustering news articles TORONTO, ONTARIO: Air France crash pilot to be quizzed 2005/08/04 BEIJING, CHINA: NK: Nuclear talks 'at a stalemate' 2005/08/04 SAN JOSE, COSTA RICA: Swimming with dolphins, sharks barred 2005/08/04 LONDON, ENGLAND: Blair in anti-hatred crackdown 2005/08/05 ADDIS ABABA, ETHIOPIA: African leaders discuss U.N. role 2005/08/04 WASHINGTON: Poll: Bush's Iraq rating at low point 2005/08/05 MOSCOW, RUSSIA: Trapped sub surfaces, crew safe 2005/08/06 Various ways to cluster articles : region, date, topic, etc. Each field has different characteristics. Text string, Time, Geographic Location, Number, etc. [ [6]

7 Clustering Algorithms Parametric Model-based Approaches Graph-based Approaches [7]

8 Graph-based Clustering Similarity = Edge Weight in a Graph Clustering = Partitioning a complete graph Spectral clustering, Normalized cut, etc. Pros and Cons No metric space assumption. Discrete or categorical fields can be handled. Ability to handle all possible partitions. Computationally expensive. [8]

9 Measuring Similarity How is the Similarity determined? Affinity in [0,1] : from totally different to identical. It depends on the type of data and the goal of clustering. (Pairwise) Affinity Matrix (i,j) element represents the similarity between data i and j. A symmetric and non-negative matrix. [9]

10 From Distance To Affintiy Often the distance is easier to measure from data. Conversion function : A D [10]

11 Spectral Clustering How to assign each point into a group? Use the eigenvectors of the normalized affinity matrix associated with the k largest eigenvalues. Run k-means algorithm on the embedded points. [Ng,Jordan,Weiss NIPS01] [11]

12 Partitioning Graphs : Normalized Cut Minimum Cut vs. Normalized Cut A B cut A B assoc [WuLeahy PAMI93] [Shi,Malik PAMI00] [12]

13 K-way Normalized Cut K-way Normalized Cut S 2 Use the eigenvectors of L associated with the 2 nd ~(k+1) th eigenvalues. Treat each row of the eigenvector matrix as a point in K. Run k-means algorithm on the embedded points. S 1 R S 3 [Bach,Jordan NIPS03] [13]

14 Multi-adic Similarity Measure What if a measure of similarity is not available for pairs of data elements, but one is available for larger groups of elements? For example, consider k-lines clustering problem. Every pair of points trivially defines a line. At least three points are needed to measure the fitting error. [14]

15 Graph vs. Hypergraph Pairwise similarity : Graph edge = Multi-adic similarity : Hyperedge The similarity value = the (hyper-)edge weight. Hyperedges can contain any number of vertices. Hypergraph A set of vertices (i.e. data elements). A set of hyperedges { z ; z = {i, j, k, l} }. D h(z) i j k l z = { i, j, k, l } A weight function h(z). [15]

16 Working with Hypergraphs Work directly on a hypergraph? It is not computationally feasible. ( # of hyperedges = ) Few analysis methods are available. Approximate with a graph. Some information will be lost. However, we only need to preserve the cluster structure. [16]

17 Approximation of a Hypergraph How to convert hyperedges to edges? Star Expansion Replace each hyperedge with one additional vertex and k edges. h(z) d c b a d c e a b g(a,e) h(z) Clique Expansion Create a clique on the vertices with edges. h(z) d c b a d c a b g(a,b) g(a,b) + h(z) [Hu,Morder VLSICircuitLayout 85] [Hadley DiscreteApplMath 95] [17]

18 Clique Averaging Problems in previous approaches Star Expansion creates too many new vertices. Clique Expansion just accumulates the hyperedge weights to the corresponding graph edges. Optimally distribute a hyperedge weight to its clique edges. h(z) d c b a d c a b [Agarwal,Lim,Zelnik,Perona,Kriegman,Belongie CVPR05, to appear] [18]

19 Clique Averaging Solve a constrained linear system of the equations. Solving where is a zero-one matrix s.t. ij is 1 if edge j is incident on hyperedge i, and 0 o.w. A highly over-determined, sparse and low-rank linear system. is huge : subsample the hypergraph. [19]

20 Hypergraph Clustering Algorithm [20]

21 Experiment: k-curves problem Synthetic Data: Cluster points sampled from k curves. Curves are in 5D space and pass the origin. Curves are gently curved and points are sampled with noise. 350 points from 5 curves. Similarity measure Pick 3 points. Compute the 2 nd singular value s 2 after subtracting the mean. [21]

22 Experiment: k-curves problem Pilot Experiment : parameter sweep Experimental Result Algorithms Error CliqueAverage 12.6 CliqueExpansion 12.9 GibsonSum 17.3 GibsonProd 55.1 KHMETIS 18.0 CRANSAC 23.4 [22]

23 Experiment: Yale Face Database Cluster images over illumination variation. The dataset contains 450 images of 10 people. Images are cropped and resized into Similarity measure Pick 4 images. Compute 4 th singular value s 4. [23]

24 Experiment: Yale Face Database Pilot Experiment : parameter sweep (7 people) [24]

25 Experiment: Yale Face Database # people CliqueAverage 4.2/ / / /3.0 CliequeExpansion 11.8/ / / /4.3 CRANSAC 16.2/ / / /6.6 KHMETIS 21.5/ / / /3.3 GibsonSum 25.9/ / / /2.1 GibsonProd 67.4/ / / /0.7 Average error rate % / standard deviation over 30 randomly sampled sub-datasets. [25]

26 Discussion: Hypergraph Clustering Multi-adic similarity measure In many problems, this is a useful affinity measure. Find an approximate graph, then cluster it. Clique Averaging It generates better and more robust approximate graphs. The relation to other algorithms is presented. [26]

27 Images of an Object Images of an object may look very different, due to External / Environmental causes When pose (or viewing direction) changes. When lighting changes. Intrinsic causes Articulation, deformation. Human face Facial expression changes. Glasses, facial hair. Aging. [27]

28 Image as a Point in the Image Space Image of size (w h) Point in (wh)-dim Space w h wh [28]

29 Image Clustering over Illumination Objects under different lightings but in a fixed pose. [Ho,Yang,Lim,Lee,Kriegman CVPR03] [29]

30 Illumination Cone Illumination Cone Theorem The set of n-pixel images of any object, seen under all possible lighting conditions, is a convex cone in n. Derived from the superposition property of illumination. An image in the cone can be written as a convex combination of other images in the cone. R = α 1 + α 2 + α 3, α 1, α 2, α 3 0 [Belhumeur,Kriegman IJCV98] [30]

31 Similarity Measure #0 Multi-adic Subspace measure The cone is approximated by a low-dimensional subspace. Images of a Lambertian object under a distant point light source without any shadow : 3D subspace. For 4 images, h(z) = e( /σ) h(z) = e( /σ) h(z) = e( /σ) [31]

32 Similarity Measure #1 Conic Affinity If I = α 1 I α k I k, α i 0, then α i (I I i ) / I i 2 and α i I / I i. Each α i encodes the similarity of I i to I. The conic affinity of an image shows its relation to other images in the set. Global measure conic linear affinity [32]

33 Gradient Images and Illumination Statistical Image Gradient Analysis Image gradient depends on both object geometry and surface reflectance property (albedo). The magnitude and orientation of image gradient form a joint probability distribution. [Chen,Belhumeur,Jacobs IJCV,CVPR00] [33]

34 Similarity Measure #2 Gradient Affinity Compare the difference in the direction and strength of the image gradient vector at each pixel in two images. Pairwise measure [34]

35 Experiments: YaleB and PIE database [Georghiades,Belhumeur,Kriegman PAMI01] [Sim,Baker,Bsat CMUTR01] [35]

36 Experiments: YaleB and PIE database Yale B Face Database 450 images of 10 people under 45 different lighting. Single directional light sources. Two poses : frontal and profile views. CMU PIE Database 1386 images of 66 people under 21 different lighting. Directional light + ambient background light. PIE20 : 10 datasets of randomly selected 20 people. [36]

37 Experiments: YaleB and PIE database Methods Yale B (frontal) Yale B (non-frontal) PIE66 (frontal) Conic Affinity 0.44 % 4.22 % 4.18 % Gradient Affinity 1.78 % 2.22 % 3.97 % L 2 + k-means % % % L 2 + Spectral % % % Minimum error rate % [37]

38 Conic Affinity Conversion Function More control on the effective affinity range Original conic affinity : A = (W + W T ) / 2 Modified affinity : A A σ = 10 σ = 100 [38]

39 Parameter Sweep Figure Error Rate Error Rate % # Nearest Neighbor log 10 (σ) [39]

40 Parameter Sweep on Conic Affinity Yale10 YalePose10 PIE20 [40]

41 Parameter Sweep on Gradient Affinity Yale10 YalePose10 PIE20 [41]

42 Image Clustering over Pose Variation Objects seen from different Viewing Directions (under same Illumination condition). [Lim,Ho,Yang,Lee,Kriegman ECCV04] [42]

43 How are the images related? Appearance Manifold Images of an object taken from a moving camera (equivalently a moving object from a fixed camera) form a complex but locally-connected structure in the image space. Geodesic distance/relation is more important than Euclidean distance. [Murase,Nayar IJCV95] [43]

44 Secant Chord Approximation x can be approximated by its neighbors y 1 and y 2. x In general, x can be approximated with y s as y 1 y 2 and weights are recovered by solving Secant Chord y 1 x y 3 y 2 [44]

45 Handling Small Pose Variation 2D Affine warp : translation, rotation, scale, skew Preserves parallelism. 6 parameters : (t x, t y, θ, s, α, φ) Why do we use the 2D affine warp? Α Appearance changes due to small 3D motions can be well approximated by 2D affine warps. = A = [45]

46 Local Linear Structure LLS Affinity Measure Each a i can be used as an affinity measure between an image x and affine-corrected y i. Only y i s close to x are considered : Local measure [46]

47 Experiments: COIL20, VEH10 COIL20 VEH10 [47]

48 Experiments: COIL100 [48]

49 Experiments: COIL Pose Variation Top: 5 degrees apart, Middle : 10 degrees apart, Boottom : 20 degrees apart [49]

50 Experiments: FACE10 Top: 10 subjects in the dataset, Bottom: typical pose variation [50]

51 Experiments: Clustering Result Datasets Error rate (%) FACE images 0.00 COIL20 COIL20.2 COIL20.4 5, 1440 images 10, 720 images 20, 360 images COIL20s 10 random, 500 images (min.avg.) VEH10.2 COIL100.2 COIL , 360 images 10, 3600 images 20, 1800 images [51]

52 Experiments: Comparison Datasets COIL20.2 COIL20.4 FACE10 VEH10.2 Proposed Algorithm 0.00 % % 0.00 % % Affine+KNN+Spectral 7.36 % % % % Affine + Spectral % % % % Euclidean + Spectral % % % % Euclidean + k-means % % % % [52]

53 Parameter Sweep on LLS COIL20.2 COIL20.4 COIL20s [53]

54 Images with Light & Pose Variation Pose and Illumination changes simultaneously. [54]

55 YCOIL dataset Yale B face database + COIL database 18 Objects on a turntable (24 poses) Yale lighting dome : 64 lights per pose Preprocessing Resize and crop tightly (64x64 pixels) Add simulated ambient light (optional) Exclude extreme lights : use subset 1~3. Flash Positions [Personal Communication with A. Georghiades at Yale univ.] [55]

56 Experiments: YCOIL05a vs. YCOIL05b Simulated Ambient Lighting No ambient light in the original YCOIL. Hard to have correct object bounding boxes. Strong cast and attached shadows. Average images with same pose over lighting variations. YCOIL05b : 0.6 * OrgImg * AmbImg. Simulated Ambient Images [56]

57 Pose-Light Datasets Simulated Ambient Light : YCOIL05a / YCOIL05b Original YCOIL dataset is collected without any ambient lighting. 300 randomly picked images of 5 randomly chosen objects. Larger Experiment : YCOIL10 With simulated ambient lights (as YCOIL05b) 500 randomly picked images of 10 randomly chosen objects. [57]

58 Geodesic Expansion of Affinity Matrix Geodesic distance Sparsely connected graph Fill the distance matrix with geodesic distances between points. Affinity Distance : D ij = 1 / A ij. Sparse Affinity Sparse Distance Full Distance Full Affinity Geodesic Distance [Image from Tenebaum et. al. Science00] [58]

59 Empirical Joint PDF on Image Gradients Build an estimate of a probability distribution from real data. Image gradient vectors at each pixel in image pairs of same object, same pose, different lighting. Yale B face database JPDF over 2 variables : angular difference, magnitude difference JPDF over 3 variables : angular difference, two magnitudes 2D Joint PDF from Yale B database [59]

60 Local Affine-corrected Gradient Affinity Affine correction on nearby images Similarity is computed by Image Gradient JPDF 2D JPDF : LGA2D 3D JPDF : LGA3D [60]

61 Experiments: YCOIL Results YCOIL05a YCOIL05b YCOIL10 LLS 15.13% 11.13% 29.93% LGA2D 34.47% 10.20% 19.24% LGA3D After Geodesic Expansion 27.50% 6.77% 17.99% Spectral on YCOIL10 : 51.92% YCOIL05a YCOIL05b YCOIL10 geodesic - LLS 12.90% 9.57% 24.36% geodesic - LGA2D 32.00% 9.93% 19.70% geodesic - LGA3D 23.40% 6.77% 17.81% [61]

62 Parameter Scan: YCOIL05a / YCOIL05b YCOIL05a LLS LGAcb2 LGAcb3 YCOIL05b [62]

63 Parameter Scan on YCOIL10 Dataset LLS LGAcb2 LGAcb3 geolls geolgacb2 geolgacb3 [63]

64 Discussion: Image Clustering Affinity measure must reflect less-variant features. Conic affinity captures global relations among images in the set. Gradient affinity is a pairwise measure based on the probabilistic distribution of the gradient vectors under changing illumination. Local Linear Structure measure uses 2D affine warped neighbor images to compute weights of secant chord approximation. Geodesic expansion of affinity improves clustering results. Histograms may be used as an approximate prob. distribution of image gradient vectors. [64]

65 Discussion: Comparing Cluster Results Clustering algorithms have free parameters. Simply reporting the best result may not be enough. Is it possible to find good parameters for next similar trial? Two approaches Run pilot experiments, fix parameters, then run real experiments. Take average of many experiments, and plot over parameters. [65]

66 Future Direction Pose-Light Image Clustering Alignment of two images. Feature/Part-based approaches. Evaluation of Clustering Result Large dataset of images. Object detection / segmentation from real-world images. [66]

67 Thank You! Hypergraph Clustering Clique Averaging Image Clustering over Illumination and Pose Variation Conic Affinity Gradient Affinity Local Linear Structure + 2D Affine Image Alignment Gradient Affinity with 2D Affine Alignment Conclusion & Future Work Image Clustering needs study on less-variant Image characteristics. Fair Comparison of Clustering Algorithms [67]

Photometric stereo. Recovering the surface f(x,y) Three Source Photometric stereo: Step1. Reflectance Map of Lambertian Surface

Photometric stereo. Recovering the surface f(x,y) Three Source Photometric stereo: Step1. Reflectance Map of Lambertian Surface Photometric stereo Illumination Cones and Uncalibrated Photometric Stereo Single viewpoint, multiple images under different lighting. 1. Arbitrary known BRDF, known lighting 2. Lambertian BRDF, known lighting

More information

Clustering Appearances of Objects Under Varying Illumination Conditions

Clustering Appearances of Objects Under Varying Illumination Conditions Clustering Appearances of Objects Under Varying Illumination Conditions Jeffrey Ho Ming-Hsuan Yang Jongwoo Lim Kuang-Chih Lee David Kriegman jho@cs.ucsd.edu myang@honda-ri.com jlim1@uiuc.edu klee10@uiuc.edu

More information

Large-Scale Face Manifold Learning

Large-Scale Face Manifold Learning Large-Scale Face Manifold Learning Sanjiv Kumar Google Research New York, NY * Joint work with A. Talwalkar, H. Rowley and M. Mohri 1 Face Manifold Learning 50 x 50 pixel faces R 2500 50 x 50 pixel random

More information

Topology-Invariant Similarity and Diffusion Geometry

Topology-Invariant Similarity and Diffusion Geometry 1 Topology-Invariant Similarity and Diffusion Geometry Lecture 7 Alexander & Michael Bronstein tosca.cs.technion.ac.il/book Numerical geometry of non-rigid shapes Stanford University, Winter 2009 Intrinsic

More information

Lecture 8: Fitting. Tuesday, Sept 25

Lecture 8: Fitting. Tuesday, Sept 25 Lecture 8: Fitting Tuesday, Sept 25 Announcements, schedule Grad student extensions Due end of term Data sets, suggestions Reminder: Midterm Tuesday 10/9 Problem set 2 out Thursday, due 10/11 Outline Review

More information

On Order-Constrained Transitive Distance

On Order-Constrained Transitive Distance On Order-Constrained Transitive Distance Author: Zhiding Yu and B. V. K. Vijaya Kumar, et al. Dept of Electrical and Computer Engineering Carnegie Mellon University 1 Clustering Problem Important Issues:

More information

Improving Shape retrieval by Spectral Matching and Meta Similarity

Improving Shape retrieval by Spectral Matching and Meta Similarity 1 / 21 Improving Shape retrieval by Spectral Matching and Meta Similarity Amir Egozi (BGU), Yosi Keller (BIU) and Hugo Guterman (BGU) Department of Electrical and Computer Engineering, Ben-Gurion University

More information

Part-based and local feature models for generic object recognition

Part-based and local feature models for generic object recognition Part-based and local feature models for generic object recognition May 28 th, 2015 Yong Jae Lee UC Davis Announcements PS2 grades up on SmartSite PS2 stats: Mean: 80.15 Standard Dev: 22.77 Vote on piazza

More information

Robust Face Recognition via Sparse Representation Authors: John Wright, Allen Y. Yang, Arvind Ganesh, S. Shankar Sastry, and Yi Ma

Robust Face Recognition via Sparse Representation Authors: John Wright, Allen Y. Yang, Arvind Ganesh, S. Shankar Sastry, and Yi Ma Robust Face Recognition via Sparse Representation Authors: John Wright, Allen Y. Yang, Arvind Ganesh, S. Shankar Sastry, and Yi Ma Presented by Hu Han Jan. 30 2014 For CSE 902 by Prof. Anil K. Jain: Selected

More information

From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose

From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 23, NO. 6, JUNE 2001 643 From Few to Many: Illumination Cone Models for Face Recognition under Variable Lighting and Pose Athinodoros

More information

Visual Representations for Machine Learning

Visual Representations for Machine Learning Visual Representations for Machine Learning Spectral Clustering and Channel Representations Lecture 1 Spectral Clustering: introduction and confusion Michael Felsberg Klas Nordberg The Spectral Clustering

More information

Announcements. Recognition I. Gradient Space (p,q) What is the reflectance map?

Announcements. Recognition I. Gradient Space (p,q) What is the reflectance map? Announcements I HW 3 due 12 noon, tomorrow. HW 4 to be posted soon recognition Lecture plan recognition for next two lectures, then video and motion. Introduction to Computer Vision CSE 152 Lecture 17

More information

Previously. Part-based and local feature models for generic object recognition. Bag-of-words model 4/20/2011

Previously. Part-based and local feature models for generic object recognition. Bag-of-words model 4/20/2011 Previously Part-based and local feature models for generic object recognition Wed, April 20 UT-Austin Discriminative classifiers Boosting Nearest neighbors Support vector machines Useful for object recognition

More information

Announcements, schedule. Lecture 8: Fitting. Weighted graph representation. Outline. Segmentation by Graph Cuts. Images as graphs

Announcements, schedule. Lecture 8: Fitting. Weighted graph representation. Outline. Segmentation by Graph Cuts. Images as graphs Announcements, schedule Lecture 8: Fitting Tuesday, Sept 25 Grad student etensions Due of term Data sets, suggestions Reminder: Midterm Tuesday 10/9 Problem set 2 out Thursday, due 10/11 Outline Review

More information

Numerical Analysis and Statistics on Tensor Parameter Spaces

Numerical Analysis and Statistics on Tensor Parameter Spaces Numerical Analysis and Statistics on Tensor Parameter Spaces SIAM - AG11 - Tensors Oct. 7, 2011 Overview Normal Mean / Karcher Mean Karcher mean / Normal mean - finding representatives for a set of points

More information

MSA220 - Statistical Learning for Big Data

MSA220 - Statistical Learning for Big Data MSA220 - Statistical Learning for Big Data Lecture 13 Rebecka Jörnsten Mathematical Sciences University of Gothenburg and Chalmers University of Technology Clustering Explorative analysis - finding groups

More information

Surface Registration. Gianpaolo Palma

Surface Registration. Gianpaolo Palma Surface Registration Gianpaolo Palma The problem 3D scanning generates multiple range images Each contain 3D points for different parts of the model in the local coordinates of the scanner Find a rigid

More information

Head Frontal-View Identification Using Extended LLE

Head Frontal-View Identification Using Extended LLE Head Frontal-View Identification Using Extended LLE Chao Wang Center for Spoken Language Understanding, Oregon Health and Science University Abstract Automatic head frontal-view identification is challenging

More information

Non-linear dimension reduction

Non-linear dimension reduction Sta306b May 23, 2011 Dimension Reduction: 1 Non-linear dimension reduction ISOMAP: Tenenbaum, de Silva & Langford (2000) Local linear embedding: Roweis & Saul (2000) Local MDS: Chen (2006) all three methods

More information

Incremental Subspace Tracking ECCV 2004

Incremental Subspace Tracking ECCV 2004 Incremental Subspace Tracking p. 1/17 Incremental Subspace Tracking ECCV 2004 David Ross, Jongwoo Lim (UCSD), Ming-Hsuan Yang (HRI) January 19, 2004 Incremental Subspace Tracking p. 2/17 Introduction an

More information

Dimension Reduction CS534

Dimension Reduction CS534 Dimension Reduction CS534 Why dimension reduction? High dimensionality large number of features E.g., documents represented by thousands of words, millions of bigrams Images represented by thousands of

More information

DISTANCE MAPS: A ROBUST ILLUMINATION PREPROCESSING FOR ACTIVE APPEARANCE MODELS

DISTANCE MAPS: A ROBUST ILLUMINATION PREPROCESSING FOR ACTIVE APPEARANCE MODELS DISTANCE MAPS: A ROBUST ILLUMINATION PREPROCESSING FOR ACTIVE APPEARANCE MODELS Sylvain Le Gallou*, Gaspard Breton*, Christophe Garcia*, Renaud Séguier** * France Telecom R&D - TECH/IRIS 4 rue du clos

More information

Training-Free, Generic Object Detection Using Locally Adaptive Regression Kernels

Training-Free, Generic Object Detection Using Locally Adaptive Regression Kernels Training-Free, Generic Object Detection Using Locally Adaptive Regression Kernels IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIENCE, VOL.32, NO.9, SEPTEMBER 2010 Hae Jong Seo, Student Member,

More information

Misalignment-Robust Face Recognition

Misalignment-Robust Face Recognition Misalignment-Robust Face Recognition Huan Wang 1 Shuicheng Yan 2 Thomas Huang 3 Jianzhuang Liu 1 Xiaoou Tang 1,4 1 IE, Chinese University 2 ECE, National University 3 ECE, University of Illinois 4 Microsoft

More information

The Analysis of Parameters t and k of LPP on Several Famous Face Databases

The Analysis of Parameters t and k of LPP on Several Famous Face Databases The Analysis of Parameters t and k of LPP on Several Famous Face Databases Sujing Wang, Na Zhang, Mingfang Sun, and Chunguang Zhou College of Computer Science and Technology, Jilin University, Changchun

More information

Segmentation Computer Vision Spring 2018, Lecture 27

Segmentation Computer Vision Spring 2018, Lecture 27 Segmentation http://www.cs.cmu.edu/~16385/ 16-385 Computer Vision Spring 218, Lecture 27 Course announcements Homework 7 is due on Sunday 6 th. - Any questions about homework 7? - How many of you have

More information

Structured Light II. Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov

Structured Light II. Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov Structured Light II Johannes Köhler Johannes.koehler@dfki.de Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov Introduction Previous lecture: Structured Light I Active Scanning Camera/emitter

More information

Image segmentation with a statistical shape prior

Image segmentation with a statistical shape prior Image segmentation with a statistical shape prior Arturo Mendoza Quispe & Caroline Petitjean October 26, 2015 Shape prior based image segmentation Foulonneau 04 Sans a priori Avec a priori Etyngier 07

More information

Beyond Pairwise Clustering

Beyond Pairwise Clustering Accepted at CVPR, 2005, San Diego, CA [DRAFT VERSION] Beyond Pairwise Clustering Sameer Agarwal 1, Jongwoo Lim 1, Lihi Zelnik-Manor 2, Pietro Perona 2, David Kriegman 1 Serge Belongie 1 1 Department of

More information

COSC160: Detection and Classification. Jeremy Bolton, PhD Assistant Teaching Professor

COSC160: Detection and Classification. Jeremy Bolton, PhD Assistant Teaching Professor COSC160: Detection and Classification Jeremy Bolton, PhD Assistant Teaching Professor Outline I. Problem I. Strategies II. Features for training III. Using spatial information? IV. Reducing dimensionality

More information

Dimensionality Reduction and Classification through PCA and LDA

Dimensionality Reduction and Classification through PCA and LDA International Journal of Computer Applications (09 8887) Dimensionality Reduction and Classification through and Telgaonkar Archana H. PG Student Department of CS and IT Dr. BAMU, Aurangabad Deshmukh Sachin

More information

Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching

Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching Correlated Correspondences [ASP*04] A Complete Registration System [HAW*08] In this session Advanced Global Matching Some practical

More information

Recognizing hand-drawn images using shape context

Recognizing hand-drawn images using shape context Recognizing hand-drawn images using shape context Gyozo Gidofalvi Department of Computer Science and Engineering University of California, San Diego La Jolla, CA 92037 gyozo@cs.ucsd.edu Abstract The objective

More information

Robust Face Recognition via Sparse Representation

Robust Face Recognition via Sparse Representation Robust Face Recognition via Sparse Representation Panqu Wang Department of Electrical and Computer Engineering University of California, San Diego La Jolla, CA 92092 pawang@ucsd.edu Can Xu Department of

More information

Model Fitting: The Hough transform II

Model Fitting: The Hough transform II Model Fitting: The Hough transform II Guido Gerig, CS6640 Image Processing, Utah Theory: See handwritten notes GG: HT-notes-GG-II.pdf Credits: S. Narasimhan, CMU, Spring 2006 15-385,-685, Link Svetlana

More information

Resolving the Generalized Bas-Relief Ambiguity by Entropy Minimization

Resolving the Generalized Bas-Relief Ambiguity by Entropy Minimization Resolving the Generalized Bas-Relief Ambiguity by Entropy Minimization Neil G. Alldrin Satya P. Mallick David J. Kriegman nalldrin@cs.ucsd.edu spmallick@vision.ucsd.edu kriegman@cs.ucsd.edu University

More information

Face detection and recognition. Detection Recognition Sally

Face detection and recognition. Detection Recognition Sally Face detection and recognition Detection Recognition Sally Face detection & recognition Viola & Jones detector Available in open CV Face recognition Eigenfaces for face recognition Metric learning identification

More information

3D Computer Vision. Structured Light II. Prof. Didier Stricker. Kaiserlautern University.

3D Computer Vision. Structured Light II. Prof. Didier Stricker. Kaiserlautern University. 3D Computer Vision Structured Light II Prof. Didier Stricker Kaiserlautern University http://ags.cs.uni-kl.de/ DFKI Deutsches Forschungszentrum für Künstliche Intelligenz http://av.dfki.de 1 Introduction

More information

Supervised texture detection in images

Supervised texture detection in images Supervised texture detection in images Branislav Mičušík and Allan Hanbury Pattern Recognition and Image Processing Group, Institute of Computer Aided Automation, Vienna University of Technology Favoritenstraße

More information

Decorrelated Local Binary Pattern for Robust Face Recognition

Decorrelated Local Binary Pattern for Robust Face Recognition International Journal of Advanced Biotechnology and Research (IJBR) ISSN 0976-2612, Online ISSN 2278 599X, Vol-7, Special Issue-Number5-July, 2016, pp1283-1291 http://www.bipublication.com Research Article

More information

Recovering 3D Facial Shape via Coupled 2D/3D Space Learning

Recovering 3D Facial Shape via Coupled 2D/3D Space Learning Recovering 3D Facial hape via Coupled 2D/3D pace Learning Annan Li 1,2, higuang han 1, ilin Chen 1, iujuan Chai 3, and Wen Gao 4,1 1 Key Lab of Intelligent Information Processing of CA, Institute of Computing

More information

Face Detection and Recognition in an Image Sequence using Eigenedginess

Face Detection and Recognition in an Image Sequence using Eigenedginess Face Detection and Recognition in an Image Sequence using Eigenedginess B S Venkatesh, S Palanivel and B Yegnanarayana Department of Computer Science and Engineering. Indian Institute of Technology, Madras

More information

Part based models for recognition. Kristen Grauman

Part based models for recognition. Kristen Grauman Part based models for recognition Kristen Grauman UT Austin Limitations of window-based models Not all objects are box-shaped Assuming specific 2d view of object Local components themselves do not necessarily

More information

Model Fitting. Introduction to Computer Vision CSE 152 Lecture 11

Model Fitting. Introduction to Computer Vision CSE 152 Lecture 11 Model Fitting CSE 152 Lecture 11 Announcements Homework 3 is due May 9, 11:59 PM Reading: Chapter 10: Grouping and Model Fitting What to do with edges? Segment linked edge chains into curve features (e.g.,

More information

Correspondence. CS 468 Geometry Processing Algorithms. Maks Ovsjanikov

Correspondence. CS 468 Geometry Processing Algorithms. Maks Ovsjanikov Shape Matching & Correspondence CS 468 Geometry Processing Algorithms Maks Ovsjanikov Wednesday, October 27 th 2010 Overall Goal Given two shapes, find correspondences between them. Overall Goal Given

More information

Detecting Multiple Symmetries with Extended SIFT

Detecting Multiple Symmetries with Extended SIFT 1 Detecting Multiple Symmetries with Extended SIFT 2 3 Anonymous ACCV submission Paper ID 388 4 5 6 7 8 9 10 11 12 13 14 15 16 Abstract. This paper describes an effective method for detecting multiple

More information

Face2Face Comparing faces with applications Patrick Pérez. Inria, Rennes 2 Oct. 2014

Face2Face Comparing faces with applications Patrick Pérez. Inria, Rennes 2 Oct. 2014 Face2Face Comparing faces with applications Patrick Pérez Inria, Rennes 2 Oct. 2014 Outline Metric learning for face comparison Expandable parts model and occlusions Face sets comparison Identity-based

More information

Monocular Template-based Reconstruction of Inextensible Surfaces

Monocular Template-based Reconstruction of Inextensible Surfaces Monocular Template-based Reconstruction of Inextensible Surfaces Mathieu Perriollat 1 Richard Hartley 2 Adrien Bartoli 1 1 LASMEA, CNRS / UBP, Clermont-Ferrand, France 2 RSISE, ANU, Canberra, Australia

More information

Homework 4: Clustering, Recommenders, Dim. Reduction, ML and Graph Mining (due November 19 th, 2014, 2:30pm, in class hard-copy please)

Homework 4: Clustering, Recommenders, Dim. Reduction, ML and Graph Mining (due November 19 th, 2014, 2:30pm, in class hard-copy please) Virginia Tech. Computer Science CS 5614 (Big) Data Management Systems Fall 2014, Prakash Homework 4: Clustering, Recommenders, Dim. Reduction, ML and Graph Mining (due November 19 th, 2014, 2:30pm, in

More information

Vision par ordinateur

Vision par ordinateur Epipolar geometry π Vision par ordinateur Underlying structure in set of matches for rigid scenes l T 1 l 2 C1 m1 l1 e1 M L2 L1 e2 Géométrie épipolaire Fundamental matrix (x rank 2 matrix) m2 C2 l2 Frédéric

More information

Learning Articulated Skeletons From Motion

Learning Articulated Skeletons From Motion Learning Articulated Skeletons From Motion Danny Tarlow University of Toronto, Machine Learning with David Ross and Richard Zemel (and Brendan Frey) August 6, 2007 Point Light Displays It's easy for humans

More information

Recognition: Face Recognition. Linda Shapiro EE/CSE 576

Recognition: Face Recognition. Linda Shapiro EE/CSE 576 Recognition: Face Recognition Linda Shapiro EE/CSE 576 1 Face recognition: once you ve detected and cropped a face, try to recognize it Detection Recognition Sally 2 Face recognition: overview Typical

More information

Spectral clustering Lecture 3

Spectral clustering Lecture 3 Spectral clustering Lecture 3 Vasileios Zografos zografos@isy.liu.se Klas Nordberg klas@isy.liu.se Contents Computer Vision Laboratory Applications of spectral clustering Mainly computer vision (image,

More information

3D Shape Recovery of Smooth Surfaces: Dropping the Fixed Viewpoint Assumption

3D Shape Recovery of Smooth Surfaces: Dropping the Fixed Viewpoint Assumption IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL., NO., 1 3D Shape Recovery of Smooth Surfaces: Dropping the Fixed Viewpoint Assumption Yael Moses Member, IEEE and Ilan Shimshoni Member,

More information

Video based Animation Synthesis with the Essential Graph. Adnane Boukhayma, Edmond Boyer MORPHEO INRIA Grenoble Rhône-Alpes

Video based Animation Synthesis with the Essential Graph. Adnane Boukhayma, Edmond Boyer MORPHEO INRIA Grenoble Rhône-Alpes Video based Animation Synthesis with the Essential Graph Adnane Boukhayma, Edmond Boyer MORPHEO INRIA Grenoble Rhône-Alpes Goal Given a set of 4D models, how to generate realistic motion from user specified

More information

Introduction to Machine Learning

Introduction to Machine Learning Introduction to Machine Learning Clustering Varun Chandola Computer Science & Engineering State University of New York at Buffalo Buffalo, NY, USA chandola@buffalo.edu Chandola@UB CSE 474/574 1 / 19 Outline

More information

Generic Face Alignment Using an Improved Active Shape Model

Generic Face Alignment Using an Improved Active Shape Model Generic Face Alignment Using an Improved Active Shape Model Liting Wang, Xiaoqing Ding, Chi Fang Electronic Engineering Department, Tsinghua University, Beijing, China {wanglt, dxq, fangchi} @ocrserv.ee.tsinghua.edu.cn

More information

CS 664 Structure and Motion. Daniel Huttenlocher

CS 664 Structure and Motion. Daniel Huttenlocher CS 664 Structure and Motion Daniel Huttenlocher Determining 3D Structure Consider set of 3D points X j seen by set of cameras with projection matrices P i Given only image coordinates x ij of each point

More information

Selecting Models from Videos for Appearance-Based Face Recognition

Selecting Models from Videos for Appearance-Based Face Recognition Selecting Models from Videos for Appearance-Based Face Recognition Abdenour Hadid and Matti Pietikäinen Machine Vision Group Infotech Oulu and Department of Electrical and Information Engineering P.O.

More information

Announcements. Photometric Stereo. Shading reveals 3-D surface geometry. Photometric Stereo Rigs: One viewpoint, changing lighting

Announcements. Photometric Stereo. Shading reveals 3-D surface geometry. Photometric Stereo Rigs: One viewpoint, changing lighting Announcements Today Photometric Stereo, next lecture return to stereo Photometric Stereo Introduction to Computer Vision CSE152 Lecture 16 Shading reveals 3-D surface geometry Two shape-from-x methods

More information

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm Group 1: Mina A. Makar Stanford University mamakar@stanford.edu Abstract In this report, we investigate the application of the Scale-Invariant

More information

Face Recognition via Sparse Representation

Face Recognition via Sparse Representation Face Recognition via Sparse Representation John Wright, Allen Y. Yang, Arvind, S. Shankar Sastry and Yi Ma IEEE Trans. PAMI, March 2008 Research About Face Face Detection Face Alignment Face Recognition

More information

Dynamic Shape Tracking via Region Matching

Dynamic Shape Tracking via Region Matching Dynamic Shape Tracking via Region Matching Ganesh Sundaramoorthi Asst. Professor of EE and AMCS KAUST (Joint work with Yanchao Yang) The Problem: Shape Tracking Given: exact object segmentation in frame1

More information

Nine Points of Light: Acquiring Subspaces for Face Recognition under Variable Lighting

Nine Points of Light: Acquiring Subspaces for Face Recognition under Variable Lighting To Appear in CVPR 2001 Nine Points of Light: Acquiring Subspaces for Face Recognition under Variable Lighting Kuang-Chih Lee Jeffrey Ho David Kriegman Beckman Institute and Computer Science Department

More information

3D Photography: Active Ranging, Structured Light, ICP

3D Photography: Active Ranging, Structured Light, ICP 3D Photography: Active Ranging, Structured Light, ICP Kalin Kolev, Marc Pollefeys Spring 2013 http://cvg.ethz.ch/teaching/2013spring/3dphoto/ Schedule (tentative) Feb 18 Feb 25 Mar 4 Mar 11 Mar 18 Mar

More information

3D Models and Matching

3D Models and Matching 3D Models and Matching representations for 3D object models particular matching techniques alignment-based systems appearance-based systems GC model of a screwdriver 1 3D Models Many different representations

More information

Lecture 2 September 3

Lecture 2 September 3 EE 381V: Large Scale Optimization Fall 2012 Lecture 2 September 3 Lecturer: Caramanis & Sanghavi Scribe: Hongbo Si, Qiaoyang Ye 2.1 Overview of the last Lecture The focus of the last lecture was to give

More information

Learning to Recognize Faces in Realistic Conditions

Learning to Recognize Faces in Realistic Conditions 000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050

More information

Last week. Multi-Frame Structure from Motion: Multi-View Stereo. Unknown camera viewpoints

Last week. Multi-Frame Structure from Motion: Multi-View Stereo. Unknown camera viewpoints Last week Multi-Frame Structure from Motion: Multi-View Stereo Unknown camera viewpoints Last week PCA Today Recognition Today Recognition Recognition problems What is it? Object detection Who is it? Recognizing

More information

Face Re-Lighting from a Single Image under Harsh Lighting Conditions

Face Re-Lighting from a Single Image under Harsh Lighting Conditions Face Re-Lighting from a Single Image under Harsh Lighting Conditions Yang Wang 1, Zicheng Liu 2, Gang Hua 3, Zhen Wen 4, Zhengyou Zhang 2, Dimitris Samaras 5 1 The Robotics Institute, Carnegie Mellon University,

More information

Multiple Motion Scene Reconstruction from Uncalibrated Views

Multiple Motion Scene Reconstruction from Uncalibrated Views Multiple Motion Scene Reconstruction from Uncalibrated Views Mei Han C & C Research Laboratories NEC USA, Inc. meihan@ccrl.sj.nec.com Takeo Kanade Robotics Institute Carnegie Mellon University tk@cs.cmu.edu

More information

Segmentation and Tracking of Partial Planar Templates

Segmentation and Tracking of Partial Planar Templates Segmentation and Tracking of Partial Planar Templates Abdelsalam Masoud William Hoff Colorado School of Mines Colorado School of Mines Golden, CO 800 Golden, CO 800 amasoud@mines.edu whoff@mines.edu Abstract

More information

Data fusion and multi-cue data matching using diffusion maps

Data fusion and multi-cue data matching using diffusion maps Data fusion and multi-cue data matching using diffusion maps Stéphane Lafon Collaborators: Raphy Coifman, Andreas Glaser, Yosi Keller, Steven Zucker (Yale University) Part of this work was supported by

More information

Subspace Clustering with Global Dimension Minimization And Application to Motion Segmentation

Subspace Clustering with Global Dimension Minimization And Application to Motion Segmentation Subspace Clustering with Global Dimension Minimization And Application to Motion Segmentation Bryan Poling University of Minnesota Joint work with Gilad Lerman University of Minnesota The Problem of Subspace

More information

Det De e t cting abnormal event n s Jaechul Kim

Det De e t cting abnormal event n s Jaechul Kim Detecting abnormal events Jaechul Kim Purpose Introduce general methodologies used in abnormality detection Deal with technical details of selected papers Abnormal events Easy to verify, but hard to describe

More information

Heat Kernel Based Local Binary Pattern for Face Representation

Heat Kernel Based Local Binary Pattern for Face Representation JOURNAL OF LATEX CLASS FILES 1 Heat Kernel Based Local Binary Pattern for Face Representation Xi Li, Weiming Hu, Zhongfei Zhang, Hanzi Wang Abstract Face classification has recently become a very hot research

More information

Final Project Face Detection and Recognition

Final Project Face Detection and Recognition Final Project Face Detection and Recognition Submission Guidelines: 1. Follow the guidelines detailed in the course website and information page.. Submission in pairs is allowed for all students registered

More information

Lambertian model of reflectance II: harmonic analysis. Ronen Basri Weizmann Institute of Science

Lambertian model of reflectance II: harmonic analysis. Ronen Basri Weizmann Institute of Science Lambertian model of reflectance II: harmonic analysis Ronen Basri Weizmann Institute of Science Illumination cone What is the set of images of an object under different lighting, with any number of sources?

More information

Nonrigid Surface Modelling. and Fast Recovery. Department of Computer Science and Engineering. Committee: Prof. Leo J. Jia and Prof. K. H.

Nonrigid Surface Modelling. and Fast Recovery. Department of Computer Science and Engineering. Committee: Prof. Leo J. Jia and Prof. K. H. Nonrigid Surface Modelling and Fast Recovery Zhu Jianke Supervisor: Prof. Michael R. Lyu Committee: Prof. Leo J. Jia and Prof. K. H. Wong Department of Computer Science and Engineering May 11, 2007 1 2

More information

Articulated Pose Estimation with Flexible Mixtures-of-Parts

Articulated Pose Estimation with Flexible Mixtures-of-Parts Articulated Pose Estimation with Flexible Mixtures-of-Parts PRESENTATION: JESSE DAVIS CS 3710 VISUAL RECOGNITION Outline Modeling Special Cases Inferences Learning Experiments Problem and Relevance Problem:

More information

Segmentation: Clustering, Graph Cut and EM

Segmentation: Clustering, Graph Cut and EM Segmentation: Clustering, Graph Cut and EM Ying Wu Electrical Engineering and Computer Science Northwestern University, Evanston, IL 60208 yingwu@northwestern.edu http://www.eecs.northwestern.edu/~yingwu

More information

on learned visual embedding patrick pérez Allegro Workshop Inria Rhônes-Alpes 22 July 2015

on learned visual embedding patrick pérez Allegro Workshop Inria Rhônes-Alpes 22 July 2015 on learned visual embedding patrick pérez Allegro Workshop Inria Rhônes-Alpes 22 July 2015 Vector visual representation Fixed-size image representation High-dim (100 100,000) Generic, unsupervised: BoW,

More information

Video Google faces. Josef Sivic, Mark Everingham, Andrew Zisserman. Visual Geometry Group University of Oxford

Video Google faces. Josef Sivic, Mark Everingham, Andrew Zisserman. Visual Geometry Group University of Oxford Video Google faces Josef Sivic, Mark Everingham, Andrew Zisserman Visual Geometry Group University of Oxford The objective Retrieve all shots in a video, e.g. a feature length film, containing a particular

More information

Image Segmentation. Srikumar Ramalingam School of Computing University of Utah. Slides borrowed from Ross Whitaker

Image Segmentation. Srikumar Ramalingam School of Computing University of Utah. Slides borrowed from Ross Whitaker Image Segmentation Srikumar Ramalingam School of Computing University of Utah Slides borrowed from Ross Whitaker Segmentation Semantic Segmentation Indoor layout estimation What is Segmentation? Partitioning

More information

Derivative Delay Embedding: Online Modeling of Streaming Time Series

Derivative Delay Embedding: Online Modeling of Streaming Time Series Derivative Delay Embedding: Online Modeling of Streaming Time Series Zhifei Zhang (PhD student), Yang Song, Wei Wang, and Hairong Qi Department of Electrical Engineering & Computer Science Outline 1. Challenges

More information

Learning a Manifold as an Atlas Supplementary Material

Learning a Manifold as an Atlas Supplementary Material Learning a Manifold as an Atlas Supplementary Material Nikolaos Pitelis Chris Russell School of EECS, Queen Mary, University of London [nikolaos.pitelis,chrisr,lourdes]@eecs.qmul.ac.uk Lourdes Agapito

More information

Multi-view stereo. Many slides adapted from S. Seitz

Multi-view stereo. Many slides adapted from S. Seitz Multi-view stereo Many slides adapted from S. Seitz Beyond two-view stereo The third eye can be used for verification Multiple-baseline stereo Pick a reference image, and slide the corresponding window

More information

Object and Action Detection from a Single Example

Object and Action Detection from a Single Example Object and Action Detection from a Single Example Peyman Milanfar* EE Department University of California, Santa Cruz *Joint work with Hae Jong Seo AFOSR Program Review, June 4-5, 29 Take a look at this:

More information

Online Learning of Probabilistic Appearance Manifolds for Video-based Recognition and Tracking

Online Learning of Probabilistic Appearance Manifolds for Video-based Recognition and Tracking Online Learning of Probabilistic Appearance Manifolds for Video-based Recognition and Tracking Kuang-Chih Lee David Kriegman Computer Science Computer Science & Engineering University of Illinois, Urbana-Champaign

More information

calibrated coordinates Linear transformation pixel coordinates

calibrated coordinates Linear transformation pixel coordinates 1 calibrated coordinates Linear transformation pixel coordinates 2 Calibration with a rig Uncalibrated epipolar geometry Ambiguities in image formation Stratified reconstruction Autocalibration with partial

More information

CSE 473/573 Computer Vision and Image Processing (CVIP) Ifeoma Nwogu. Lectures 21 & 22 Segmentation and clustering

CSE 473/573 Computer Vision and Image Processing (CVIP) Ifeoma Nwogu. Lectures 21 & 22 Segmentation and clustering CSE 473/573 Computer Vision and Image Processing (CVIP) Ifeoma Nwogu Lectures 21 & 22 Segmentation and clustering 1 Schedule Last class We started on segmentation Today Segmentation continued Readings

More information

Announcement. Photometric Stereo. Computer Vision I. Shading reveals 3-D surface geometry. Shape-from-X. CSE252A Lecture 8

Announcement. Photometric Stereo. Computer Vision I. Shading reveals 3-D surface geometry. Shape-from-X. CSE252A Lecture 8 Announcement Photometric Stereo Lecture 8 Read Chapter 2 of Forsyth & Ponce Office hours tomorrow: 3-5, CSE 4127 5-6, CSE B260A Piazza Next lecture Shape-from-X Shading reveals 3-D surface geometry Where

More information

Algorithms for 3D Isometric Shape Correspondence

Algorithms for 3D Isometric Shape Correspondence Algorithms for 3D Isometric Shape Correspondence Yusuf Sahillioğlu Computer Eng. Dept., Koç University, Istanbul, Turkey (PhD) Computer Eng. Dept., METU, Ankara, Turkey (Asst. Prof.) 2 / 53 Problem Definition

More information

Unsupervised discovery of category and object models. The task

Unsupervised discovery of category and object models. The task Unsupervised discovery of category and object models Martial Hebert The task 1 Common ingredients 1. Generate candidate segments 2. Estimate similarity between candidate segments 3. Prune resulting (implicit)

More information

Shape Classification and Cell Movement in 3D Matrix Tutorial (Part I)

Shape Classification and Cell Movement in 3D Matrix Tutorial (Part I) Shape Classification and Cell Movement in 3D Matrix Tutorial (Part I) Fred Park UCI icamp 2011 Outline 1. Motivation and Shape Definition 2. Shape Descriptors 3. Classification 4. Applications: Shape Matching,

More information

arxiv: v1 [cs.cv] 9 Sep 2013

arxiv: v1 [cs.cv] 9 Sep 2013 Learning Transformations for Clustering and Classification arxiv:139.274v1 [cs.cv] 9 Sep 213 Qiang Qiu Department of Electrical and Computer Engineering Duke University Durham, NC 2778, USA Guillermo Sapiro

More information

A Study on Similarity Computations in Template Matching Technique for Identity Verification

A Study on Similarity Computations in Template Matching Technique for Identity Verification A Study on Similarity Computations in Template Matching Technique for Identity Verification Lam, S. K., Yeong, C. Y., Yew, C. T., Chai, W. S., Suandi, S. A. Intelligent Biometric Group, School of Electrical

More information

Subspace Clustering. Weiwei Feng. December 11, 2015

Subspace Clustering. Weiwei Feng. December 11, 2015 Subspace Clustering Weiwei Feng December 11, 2015 Abstract Data structure analysis is an important basis of machine learning and data science, which is now widely used in computational visualization problems,

More information

3D Photography: Stereo

3D Photography: Stereo 3D Photography: Stereo Marc Pollefeys, Torsten Sattler Spring 2016 http://www.cvg.ethz.ch/teaching/3dvision/ 3D Modeling with Depth Sensors Today s class Obtaining depth maps / range images unstructured

More information

Cluster Analysis. Mu-Chun Su. Department of Computer Science and Information Engineering National Central University 2003/3/11 1

Cluster Analysis. Mu-Chun Su. Department of Computer Science and Information Engineering National Central University 2003/3/11 1 Cluster Analysis Mu-Chun Su Department of Computer Science and Information Engineering National Central University 2003/3/11 1 Introduction Cluster analysis is the formal study of algorithms and methods

More information