Object Segmentation Presented by Sherin Aly 1. What is a ‘Good Segmentation’?

Slides:



Advertisements
Similar presentations
Automatic Photo Pop-up Derek Hoiem Alexei A.Efros Martial Hebert Carnegie Mellon University.
Advertisements

Top-Down & Bottom-Up Segmentation
Top-Down & Bottom Up Segmentation
Ch2 Data Preprocessing part3 Dr. Bernard Chen Ph.D. University of Central Arkansas Fall 2009.
Image classification Given the bag-of-features representations of images from different classes, how do we learn a model for distinguishing them?
Foreground Focus: Finding Meaningful Features in Unlabeled Images Yong Jae Lee and Kristen Grauman University of Texas at Austin.
Interactively Co-segmentating Topically Related Images with Intelligent Scribble Guidance Dhruv Batra, Carnegie Mellon University Adarsh Kowdle, Cornell.
Learning to Combine Bottom-Up and Top-Down Segmentation Anat Levin and Yair Weiss School of CS&Eng, The Hebrew University of Jerusalem, Israel.
Shape Sharing for Object Segmentation
Recovering Human Body Configurations: Combining Segmentation and Recognition Greg Mori, Xiaofeng Ren, and Jitentendra Malik (UC Berkeley) Alexei A. Efros.
Carolina Galleguillos, Brian McFee, Serge Belongie, Gert Lanckriet Computer Science and Engineering Department Electrical and Computer Engineering Department.
Qualifying Exam: Contour Grouping Vida Movahedi Supervisor: James Elder Supervisory Committee: Minas Spetsakis, Jeff Edmonds York University Summer 2009.
Texture Segmentation Based on Voting of Blocks, Bayesian Flooding and Region Merging C. Panagiotakis (1), I. Grinias (2) and G. Tziritas (3)
Ghunhui Gu, Joseph J. Lim, Pablo Arbeláez, Jitendra Malik University of California at Berkeley Berkeley, CA
Object Detection by Matching Longin Jan Latecki. Contour-based object detection Database shapes: …..
Discriminative and generative methods for bags of features
Lecture 6 Image Segmentation
Recognition using Regions CVPR Outline Introduction Overview of the Approach Experimental Results Conclusion.
EE 7730 Image Segmentation.
Computer Vision Group University of California Berkeley 1 Learning Scale-Invariant Contour Completion Xiaofeng Ren, Charless Fowlkes and Jitendra Malik.
Real-time Embedded Face Recognition for Smart Home Fei Zuo, Student Member, IEEE, Peter H. N. de With, Senior Member, IEEE.
1 Learning to Detect Objects in Images via a Sparse, Part-Based Representation S. Agarwal, A. Awan and D. Roth IEEE Transactions on Pattern Analysis and.
Fast, Multiscale Image Segmentation: From Pixels to Semantics Ronen Basri The Weizmann Institute of Science Joint work with Achi Brandt, Meirav Galun,
Region Segmentation. Find sets of pixels, such that All pixels in region i satisfy some constraint of similarity.
Learning to Detect Natural Image Boundaries Using Local Brightness, Color, and Texture Cues David R. Martin Charless C. Fowlkes Jitendra Malik.
Abstract We present a model of curvilinear grouping using piecewise linear representations of contours and a conditional random field to capture continuity.
CVR05 University of California Berkeley 1 Familiar Configuration Enables Figure/Ground Assignment in Natural Scenes Xiaofeng Ren, Charless Fowlkes, Jitendra.
Berkeley Vision GroupNIPS Vancouver Learning to Detect Natural Image Boundaries Using Local Brightness,
A Database of Human Segmented Natural Images and Two Applications David Martin, Charless Fowlkes, Doron Tal, Jitendra Malik UC Berkeley
1 Segmentation with Scene and Sub-Scene Categories Joseph Djugash Input Image Scene/Sub-Scene Classification Segmentation.
Stereo Computation using Iterative Graph-Cuts
1 How do ideas from perceptual organization relate to natural scenes?
1 Ecological Statistics and Perceptual Organization Charless Fowlkes work with David Martin and Jitendra Malik at University of California at Berkeley.
Computer Vision Group University of California Berkeley 1 Cue Integration in Figure/Ground Labeling Xiaofeng Ren, Charless Fowlkes and Jitendra Malik.
Presentation By Michael Tao and Patrick Virtue. Agenda History of the problem Graph cut background Compute graph cut Extensions State of the art Continued.
Perceptual Organization: Segmentation and Optical Flow.
Cue Integration in Figure/Ground Labeling Xiaofeng Ren, Charless Fowlkes and Jitendra Malik, U.C. Berkeley We present a model of edge and region grouping.
Heather Dunlop : Advanced Perception January 25, 2006
Computer Vision - A Modern Approach Set: Segmentation Slides by D.A. Forsyth Segmentation and Grouping Motivation: not information is evidence Obtain a.
Image Segmentation Image segmentation is the operation of partitioning an image into a collection of connected sets of pixels. 1. into regions, which usually.
Recognition using Regions (Demo) Sudheendra V. Outline Generating multiple segmentations –Normalized cuts [Ren & Malik (2003)] Uniform regions –Watershed.
Multiscale Symmetric Part Detection and Grouping Alex Levinshtein, Sven Dickinson, University of Toronto and Cristian Sminchisescu, University of Bonn.
Digital Image Processing In The Name Of God Digital Image Processing Lecture8: Image Segmentation M. Ghelich Oghli By: M. Ghelich Oghli
Chapter 14: SEGMENTATION BY CLUSTERING 1. 2 Outline Introduction Human Vision & Gestalt Properties Applications – Background Subtraction – Shot Boundary.
Detecting Curved Symmetric Parts using a Deformable Disc Model Tom Sie Ho Lee, University of Toronto Sanja Fidler, TTI Chicago Sven Dickinson, University.
Efficient Subwindow Search: A Branch and Bound Framework for Object Localization ‘PAMI09 Beyond Sliding Windows: Object Localization by Efficient Subwindow.
Today Ensemble Methods. Recap of the course. Classifier Fusion
Chapter 4: Pattern Recognition. Classification is a process that assigns a label to an object according to some representation of the object’s properties.
CS654: Digital Image Analysis
Geodesic Saliency Using Background Priors
1Ellen L. Walker Category Recognition Associating information extracted from images with categories (classes) of objects Requires prior knowledge about.
Category Independent Region Proposals Ian Endres and Derek Hoiem University of Illinois at Urbana-Champaign.
Random Forests Ujjwol Subedi. Introduction What is Random Tree? ◦ Is a tree constructed randomly from a set of possible trees having K random features.
A New Method for Crater Detection Heather Dunlop November 2, 2006.
Object Recognition as Ranking Holistic Figure-Ground Hypotheses Fuxin Li and Joao Carreira and Cristian Sminchisescu 1.
Object Recognition by Integrating Multiple Image Segmentations Caroline Pantofaru, Cordelia Schmid, Martial Hebert ECCV 2008 E.
776 Computer Vision Jan-Michael Frahm Spring 2012.
Machine Vision Edge Detection Techniques ENT 273 Lecture 6 Hema C.R.
Color Image Segmentation Mentor : Dr. Rajeev Srivastava Students: Achit Kumar Ojha Aseem Kumar Akshay Tyagi.
Rich feature hierarchies for accurate object detection and semantic segmentation 2014 IEEE Conference on Computer Vision and Pattern Recognition Ross Girshick,
Tree and Forest Classification and Regression Tree Bagging of trees Boosting trees Random Forest.
Nonparametric Semantic Segmentation
Object detection as supervised classification
Brain Hemorrhage Detection and Classification Steps
Learning to Combine Bottom-Up and Top-Down Segmentation
“The Truth About Cats And Dogs”
Design of Hierarchical Classifiers for Efficient and Accurate Pattern Classification M N S S K Pavan Kumar Advisor : Dr. C. V. Jawahar.
Grouping/Segmentation
Text Categorization Berlin Chen 2003 Reference:
“Traditional” image segmentation
Presentation transcript:

Object Segmentation Presented by Sherin Aly 1

What is a ‘Good Segmentation’?

g/resources.html

Learning a classification model for segmentation Xiaofeng Ren and Jitendra Malik 4

methodology Two-class classification model Over segmentation as preprocessing They use classical Gestalt cues –Contour, texture, brightness and continuation A linear classifier is used for training 5

Good Vs Bad segmentation 6 a) Image from Corel Imagebase b) superimposed with a human marked segmentation c) Same image with Bad segmentation

How do we distinguish good segmentations from bad segmentations? 7

How? Use “Classical Gestalt cues” – proximity, similarity and good continuation Instead of Ad-hoc decision about features combination 8

Gestalt Principles of Grouping 9 In order to interpret what we receive through our senses,we attempt to organize this information into certain groups.

Methodology Preprocessing Feature extraction Feature evaluation Training Optimization Find good segmentaion 10

Preprocessing 11 Superpixel map K=200 Reconstruction of human segmentation from Superpixels a contour-based measure is used to quantify this approximation Local Coherent Preserve structure Contour texture

12 The percentage of human marked boundaries covered by the superpixel maps Tolerance 1,2,and 3

Feature Extraction 1. inter-region texture similarity 2. intra-region texture similarity 3. inter-region brightness similarity 4. intra-region brightness similarity 5. inter-region contour energy 6. intra-region contour energy 7. curvilinear continuity 13

Feature Extraction 1. inter-region texture similarity 2. intra-region texture similarity 3. inter-region brightness similarity 4. intra-region brightness similarity 5. inter-region contour energy 6. intra-region contour energy 7. curvilinear continuity 14

Feature Extraction 1. inter-region texture similarity 2. intra-region texture similarity 3. inter-region brightness similarity 4. intra-region brightness similarity 5. inter-region contour energy 6. intra-region contour energy 7. curvilinear continuity 15

Power of Gestalt cues 16 =

Training the classifier simple logistic regression classifier, 17 Empirical distribution of pairs of features

18 Precision is the fraction of detections which are true positives. Recall is the fraction of true positives which are detected

Conclusion There simple linear classifier had promising results on a variety of natural images. boundary contour is the most informative grouping cue, and it is in essence discriminative. 19

Pros & Cons Cons –The larger spatial support that superpixels provide, allowing more global features to be computed than on pixels alone. –The use of superpixels improves the computational efficiency –SuperPixels technique is very applicable Pros –Might fall in Local Minima 20

Combining Top-down and Bottom-up Segmentation Eran Borenstein Eitan Sharon Shimon Ullman 21

Motivation Bottom-Up segmentation –Rely on continuity principle –Capture image properties “texture, grey level uniformity and contour continuity” –Segmentation based on similarities between image regions How can we capture prior knowledge of a specific object (class)? –Answer: Top-Down Segmentation –use prior knowledge about an object Credit: Joseph Djugash

Bottom-Up Segmentation Slides from Eitan Sharon, “ Segmentation and Boundary Detection Using Multiscale Intensity Measurements ”. Credit: Joseph Djugash

Normalized-Cut Measure Slides from Eitan Sharon, “ Segmentation and Boundary Detection Using Multiscale Intensity Measurements ”. Credit: Joseph Djugash

Top-Down approach Input Fragments MatchingCover Credit: Joseph Djugash

Another step towards the middle Bottom-Up Top-Down Credit: Joseph Djugash

Some Definitions & Constraints Measure of saliency h(Γ i ), h i є [0,1) A configuration vector s contains labels s i (1/- 1) of all the segments (S i ) in the tree The label s i can be different from its parent’s label s i – Cost function for a given s Top-down termBottom-up term Defines the weighted edge between S i & S i –

Classification Costs The terminal segments of the tree determine the final classification The top-down term is defined as: The saliency of a segment should restrict its label (based on its parent’s label) The bottom-up term is defined as:

Minimizing the Costs – Information Exchange in a Tree Bottom-up message: Top-down message: Min-cost Label: Cost of s i = –1 and s = x Message from s i = –1 Cost of s i = +1 and s = x Message from s i = +1 Computed at each node – minimal of the values is the selected label of node s in s Minimal Cost if the region was classified as background Minimal Cost if the region was classified as figure

Confidence Map Evaluating the confidence of a region: Causes of Uncertainty of Classification –Bottom-up uncertainty – regions where there is no salient bottom-up segment matching the top-down classification –Top-down uncertainty – regions where the top- down classification is ambiguous (highly variable shape regions) The type of uncertainty and the confidence values can be used to select appropriate additional processing to improve segmentation

Results Calculate average distance between a given segmentation contour and a benchmark contour. Removing from the average all contour points having a confidence measure less than 0.1. The resulting confidence map efficiently separated regions of high and low consistency. The combined scheme improved the top-down contour by over 67% on average. This improvement was even larger in object parts with highly variable shape. 31

Results (cont.) top-down process may produce a figure-ground approximation that does not follow the image discontinuities. Salient bottom-up segments can correct these errors and delineate precise region boundaries Buttom up The initial classification map T(x, y)

Results III (cont.)

the top-down completely misses a part of the object. The confidence map may be helpful in identifying such cases,

Results III (cont.) bottom-up segmentation may be insufficient in detecting the figure-ground contour, and the top-down process completes the missing information

Results III (cont.)

Salient bottom-up segments can correct these errors and delineate precise region boundaries

Conclusion Buttom-up and top-down merits Provide reliable confidence map It take into account all discontinuities at all scales But: If the object is assigned a given category, the specific features cannot be adopted for other categories 38

Constrained Parametric Min-Cuts for Automatic Object Segmentation Joao Carreira Cristian Sminchisescu 39

Traditional Segmentation: Finding Homogeneous Regions 40 gPb-owt-ucm: P. Arbelaez, M. Maire, C. Fowlkes, and J. Malik. PAMI 2010.

Conventional Bottom-up Segmentation Proposed approach 1.Split multiple times 2.Retain object-like segmentations Bottom-up Object Segmentation Credit: J. Carreira High redundancy

Bottom-up Object Segmentation 42 Credit: J. Carreira A single multi-region segmentation or a hierarchy

Proposed Bottom-up Object Segmentation 43 Credit: J. Carreira single-shot multi- region segmentation robust set of overlapping figure-ground segmentations Segments with object-like regularities superpixels

44 Constrained Parametric Min-Cuts for Automatic Object Segmentation Credit: J. Carreira parametric max-flow solver Figure ground segmentation by growing regions around seeds Ranking

45 Constrained Parametric Min-Cuts for Automatic Object Segmentation Credit: J. Carreira

Initialization Foreground –Regular 5x5 grid geometry –Centroids of large N-Cuts regions –Centroids of superpixels closest to grid positions Background –Full image boundary –Horizontal boundaries –Vertical boundaries –All boundaries excluding the bottom one Performance broadly invariant to different initializations

Generating a segment pool: c onstrained min-cut min cut hard constraint background object hard constraint 47 Credit: J. Carreira

Generating a Segment Pool: C onstrained Parametric Min-Cuts 48 Credit: J. Carreira

49 Generating a Segment Pool: C onstrained Parametric Min-Cuts Credit: J. Carreira

50 Generating a Segment Pool: C onstrained Parametric Min-Cuts Credit: J. Carreira

Can solve for all values of object bias in the same time complexity of solving a single min-cut using a parametric max-flow solver 51 Generating a Segment Pool: C onstrained Parametric Min-Cuts Credit: J. Carreira

Fast Rejection Large set of initial segmentations (~5500) High Energy Low Energy ~2000 segments with the lowest energy Cluster segments based on spatial overlap (at least 0.95) Lowest energy member of each cluster (~154 in PASCAL VOC) Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

53 Constrained Parametric Min-Cuts for Automatic Object Segmentation Credit: J. Carreira ranks all the sampled object segmentations discard all but a small subset of confident ones.

Ranking object hypotheses mid-level, category independent f eatures  Boundary – normalized boundary energy  Region – location, perimeter, area, Euler number, orientation, contrast with background  Gestalt – convexity, smoothness Good Low boundary energy Non smooth. High Euler number High boundary energy Smooth. Euler number = 0 Bad 54 Credit: J. Carreira

Traditional Classification-based Learning 55 Credit: J. Carreira

Traditional Classification-based Learning 56 Credit: J. Carreira

Proposed Ranking-based Learning 57 Credit: J. Carreira

Segment Ranking Model data using a host of features –Graph partition properties –Region properties –Gestalt properties Apply Features Normalization Train regressor with the largest overlap ground-truth segment using Random Forests Diversify similar rankings using Maximal Marginal Relevance (MMR)

Graph Partition Properties Cut – Sum of affinities along segment boundary Ratio Cut – Sum along boundary divided by the number Normalized Cut – Sum of cut and affinity in foreground and background Unbalanced N-cut – N-cut divided by foreground affinity Thresholded boundary fraction of a cut

Region Properties Area Perimeter Relative Centroid Bounding Box properties Fitting Ellipse properties Eccentricity Orientation Convex Area Euler Number Diameter of Circle with the same area of the segment Percentage of bounding box covered Absolute distance to the center of the image

Gestalt Properties Inter-region texton similarity Intra-region texton similarity Inter-region brightness similarity Intra-region brightness similarity Inter-region contour energy Intra-region contour energy Curvilinear continuity Convexity – Ratio of foreground area to convex hull area

Feature Importance for the Random Forest regressor

Feature Importance

How to Model Segment Quality ? Best overlap with a ground truth object computed by intersection-over-union. 64 Credit: J. Carreira

What has been modeled? Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Diversifying the Ranking Diversified Original Best two hypotheses Middle two hypotheses Worst two hypotheses Segment Ranking using Maximum Marginal Relevance 66

Databases –Weizmann database F-measure criterion –MSR-Cambridge database & Pascal VOC2009 Segmentation covering

Performance Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Test of the algorithm Berkeley segmentation dataset –Complete pool of images collected –Ranked using the ranking methodology –Top ranks evaluated to test the ranking procedure How well does the algorithm perform? Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Berkeley Database Rank 269! Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Berkeley Database Rank 142! Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Berkeley Database Rank 98! Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Berkeley Database Compute the Segment Covering score for the top 40 segments of each image in the database DatabaseSegment Covering Score (Top 40) BSDS0.52 MSR Cambridge0.77 Pascal VOC0.63 DatabaseSegment Covering Score (All segments) BSDS0.61 MSR Cambridge0.85 Pascal VOC0.78

Conclusion Does Constrained Parametric Min-Cuts work well? –Yes Does Fast Rejection work well? –Yes Does Segment Ranking work well? –I don’t think so Credit: SasiKanth Bendapudi Yogeshwar Nagaraj

Image Ground-truth objects Best in segment pool Best in top- ranked CPMC Segmentation Examples

CPMC Results gPb-owt-ucm: P. Arbelaez, M. Maire, C. Fowlkes, and J. Malik. PAMI Equals the state-of-the-art on the VOC 2009 dataset using just 7 segments. 76 VOC2009CoveringNumber of Segments CPMC gPb-owt-ucm MSRCCoveringNumber of Segments CPMC gPb-owt-ucm

Ranking Diversification Diversified Original First two hypothe ses Middle two hypothes es Last two hypothes es 77 Credit:Carreira

78 Credit: J. Carreira

Ranking 79 Credit: J. Carreira

Running Demos Methodologies employed –Kmeans using: Texture RGB Texture + RGB RGB + HSV Texture + Lab + HSV 80

Running Demos Data set used –Microsoft Research Cambridge Object Recognition Image Database, version 1.0. –Used: 7 classes with 23 per class Animal-grass Trees-sky-grass Buildings-sky-grass Airplanes-sky-grass Animal-grass Faces-BG Car-wall-ground 81

Experiment Results FeaturesTextureTexture + RGB RGBRGB +HSVTexture+Lab+ HSV Animal-grass72.7%74.1% 72.3%72.6%74.1% Trees-sky- grass 37.1% 40.7%38.2%37.1% Buildings-sky- grass 44.6%42.8% 51.9% 45.4% 44.7% Airplanes-sky- grass 58.8% 54.6% 59.7% 58.7% Animal-grass64.8% 69.3% 71% 64.9% Faces-BG 100% Car-wall-ground 67.2% 68.4% 64.9% 67.2% Mean63.6%63.5%65.3%64.6%63.8% 82

Experiment Results FeaturesTextureTexture + RGB RGBRGB +HSVTexture+Lab+ HSV One iteration Elapsed time is 7.42 secs secs secs 1.5 secs7.84 sec Overall Elabsed time for experiment 19.9 mins 32.9 mins 4.4 mins4 min21 mins 83 Microsoft Research Cambridge Object Recognition Image Database, version 1.0.

84

85

86

87

Acknowledgment Dr. Devi Parikh Dr. Joao Carreira 88