In this lecture, we are going to talk about image segmentation, essentially defined as methods for grouping pixels together.

Size: px
Start display at page:

Download "In this lecture, we are going to talk about image segmentation, essentially defined as methods for grouping pixels together."

Transcription

1 In this lecture, we are going to talk about image segmentation, essentially defined as methods for grouping pixels together. We will first define the segmentation problem, overview some basic ideas of grouping in human vision, and go over several segmentation methods including mean-shift, graph-cuts, energy-based methods, and conclude with reviewing a few semantic segmentation works.

2 What is segmentation? In its most basic definition, segmentation groups some pixels in an image and form bigger entities/components out of them. These entities/components/groups are called segments.

3 All pixels in one group share some property. What this property is? That is up to the definition of grouping criteria (i.e., which pixels should be grouped together and which ones should go into different groups). There are numerous methods for it: grouping based on color, texture, semantic meaning, etc.

4 This criteria can be based on different levels of abstraction. For instance, you may want to group pixels together that are just sharing the same texture (low-level of abstraction). Similarly, one may want to group pixels that share the same semantic meaning (e.g., all pixels on a dog) regardless of their diverse texture or color (higher level of abstraction).

5 As an extreme example, let s look at this image. One may want to segment out the big face, while one may want to segment the image into, say 8, regions that just share a similar color. That way, the big face would be decomposed into 3 or 4 regions.

6 Therefore, segmentation can happen at different levels of abstraction and even doesn t have to be on a 2D image. For instance, what you see on the right is a LiDAR scan where the object instances are segmented. In this case, the input is actually a 3D scan and not a 2D image. The bottom example shows a case where a surface model is segmented into its parts, defined based on the body parts of the fourlegged animal.

7 Now we look at the general approaches to segmentation. Generally, there are two main categories of methods. The first one, bottom-up segmentation, groups pixels together because they are locally coherent (based a defined coherency metric). The segments acquired from this approach usually show a lower level of abstraction and are commonly just coherent image regions.

8 Top-down segmentation is the opposite end of the spectrum. It groups pixels together based on their semantics. The resulting segments could actually show a diverse appearance and may even be in conflict with what bottom-up segmentation would yield (e.g., the guy in the picture who is wearing a green-ish shirt might be segmented together with the background using a bottom-up segmentation while in top-down, it would be expected to have a separate segment).

9 Why would we want to segment an image at all? There are several reasons, but here are a few: Segments are usually building blocks of many vision systems. We usually prefer to apply an operation on a segment rather than a bunch of independent pixels. It s also beneficial in terms of complexity. If one didn t have segments, the desired operation would have to be applied on pixels or some (sliding) windows, that are blind wrt the image content. Segments are essentially pixels that are similar to each other, so without losing virtually anything, the complexity would be reduced by the order of number of pixels included in the segment. Also, segments can extract the precise boundary of an object, so they provide more information compared to bounding boxes or similar coarse crops.

10 Let s have a brief discussion on how our brain performs visual grouping.

11 Gestaltism is a theory in psychology. Gestalt is a German word meaning whole or form. Gestaltism advocates that the brain is holistic and performs grouping to extract a whole out of pieces.

12 Gestalt includes a number of properties that may result in grouping some pieces together. The examples shown here (each row is one example) demonstrate cases that are identical except in one property. That property would lead to grouping some of these pieces together. For instance, the second row would have 3 groups, since there can be seen 3 groups each with 2 pieces which are in proximity of one another. As another example, the 5th case shows 6 pieces that are identical except for the direction of their motion. Therefore, each motion direction would form one group since they share their fate. The other two main properties shown here are similarity (3rd and 4th examples) and falling in a common region (the last two examples).

13 Our brain can also perform grouping based on occlusion. For instance, here we clearly see two things: one (gray) is occluding something else (black), though none of them look like any particular object we have seen before.

14 When we remove the occluder, the black object looks quite meaningless and would be less likely that our brain would group its pieces together. Therefore, the occluder had a key role in leading our brain to think there is one black object in the back.

15 The completion can happen by invisible objects too. For instance, all these cases look like a feasible object, while they are formed of disjoint pieces. What glues them together is basically an object that we don t even see, yet our brain performs the grouping.

16 Grouping can also happen based on assigning some pieces to the figure and the rest to the ground. Which piece goes to the figure and which ones to the ground can happen based on low-level bottom-up cues (just local coherency) or high level recognition (e.g. seeing a semantic object).

17 Now let s look at this case of emergence. What do you see here? A tree? A road intersection? A dog? Nothing?

18 You may have seen different things, but what matters is that we all saw something. Most of the people see a dog (shown on the right), while if we looked at the pieces forming the dog locally, they re actually quite meaningless. What this is showing is that our brain essentially aggregated meaningless pieces and perceived a bigger meaningful whole in them, according to the things it often sees.

19 Now, let s see how we can segment an image algorithmically. How would you do it? Take a minute and think Using clustering? Using edges/boundaries in the image? Using other cues/modalities like depth?...

20 Clustering is likely to be the first idea. As a very basic and fast overview, what clustering does is: given a set of data point and a feature associated with each, it groups the data points in a way that each one belongs to one cluster/group. In the example shown here, the feature is 2 dimensional and there are 4 clusters among the data points.

21 We can cast the problem of segmentation as clustering, since they essentially do something very similar: if each data point is one pixel, clustering the data points would give us coherent pixel groups.

22 You must be familiar with k-means clustering, a common method for clustering, from previous courses. Here we will discuss mean-shift. The main properties of k-means is that: it requires specifying the number of clusters and makes the assumption that the clusters have a certain shape (spherical). The primary characteristics of mean-shift is that it doesn t require to specify the number of clusters and makes no assumption about cluster shapes.

23 Now let s see how mean-shift works.

24 Mean-shift is a generic clustering technique. It can be easily used for image segmentation. Here you can see a few examples.

25 The key idea of mean-shift is that it looks for modes of density in a given distribution. Let s see that using an example.

26 Here we see a set of data points (let s say pixels) in a 2 dimensional space. There can be seen an obvious dense center in this distribution.

27 Mean-shift first starts at an arbitrary data point and looks at its neighborhood.

28 Then it finds the center of mass (mean) of the data points in this neighborhood.

29 Then the center is moved from the previous arbitrary data point to the center of mass.

30 This process continues till it converges.

31 The equation on the right shows how the center of mass (m) is found around a given data point (x). N(x) is the neighborhood around x, and K(.) is a weighting function or kernel that adjusts the contribution of each datapoint according to its distance from x. This equation is basically just a weighted mean. x m(x) says that at each iteration, x is updated to m(x) which is the new center of mass.

32 This procedure is again shown in these animated slides. We start from one point, find the center of mass, and move to it. 32

33 33

34 Again, we find the center of mass, and move to it. 34

35 35

36 And again. 36

37 37

38 And again...now we ve converged and found the mode of density. 38

39 This process will be initiated at a large number of arbitrary locations (or could be initiated at each data point). All windows that end up converging to the same peak will be merged. 39

40 Think of starting this mean-shift search process at each data point. Many of them obviously end up over the same peak. Those points are all on the same Attraction Basin. Each mode has it s own attraction basin, and all data points on one attraction basin form one cluster.

41 Here you can see the input data points to mean-shift (left), the found clusters (on the right), and the attraction basin and trajectories leading to them on the 3D plot (bottom) where the z dimension shows the density. Observe how attraction basins formed clusters that are consistent with the density of the data points.

42 Now, if we want to do image segmentation using mean-shift, we should follow very similar steps. First we define some features (color, texture, etc), extract the feature for each pixel, and follow the described mean-shift process to find the segments (i.e., clusters).

43 Here are some results. It works a fairly well bottom-up segmentation.

44 Some more results.

45 The advantages of mean-shift are: not requiring the number of clusters a priori, having few parameters to set, being robust to outliers, and not being limited to spherical clusters. The main disadvantages are: requiring kernel parameters and window size (though this is natural since mean-shift doesn t require inputting the number of clustering, the user has to specify some criteria for grouping), being computationally expensive, and not scaling well when the input feature is high dimensional.

46 Now, let s look at another way of performing clustering: graph based methods.

47 We represent the image as a graph and cut the graph into subgraphs in a way that the subgraphs correspond to image segments. The input graph should capture the similarity among pixels, so by cutting the low-similarity edges and preserving the high-similarity ones, proper segments can be found.

48 In the input graph, each node represents one pixel, and there is an edge between every pair of pixels (though sometimes we only connect pixels that are spatially nearby -- mostly to reduce complexity). The weight of each edge captures how similar the nodes/pixels are. We should mention that in the next few slides, we will discuss binary graph partitioning (i.e., the graph is cut into two disjoint pieces). Therefore, it yields two segments. There are ways for going from a binary partitioning/segmentation to multi-segment segmentation. The simplest one is just iteratively applying the same binary segmentation method on the segments found by the previous iteration. You ll try this in the homework.

49 There are different ways the similarity metric used in the edge weights can be defined. Here are some common ones: Spatial Distance: the closer the pixels, the lesser the distance between them. Intensity: the more similar the density, the lesser the distance. Color: the more similar the color, the lesser the distance. The total distance could be a (often linear) mixture of these components.

50 Now we cut this graph to find the sub-graphs. Each sub-graph represents one segment. Cutting means breaking the graph into disjoint components; cutting happens by removing edges. Obviously, we need to remove the edges with low-affinity and preserve the ones with high-affinity to get proper segments.

51 Graph-cut is a technique for finding the cut given an input graph.

52 We need to assign a cost to a feasible cut in order to find the best cut. The cost of a cut is defined as the summation of the weights of the edges the cut will remove. Based on this definition, what s the cost of the blue cut?.... It s 2+3=5.

53 As mentioned before, the problem of segmenting the image into two segments is now equivalent to finding the optimal cut on the input graph. There are good methods for solving this efficiently. However, this has a fatal bias: the cut with the lowest cost usually just cuts out one or few isolated nodes. That s because those nodes are so isolated and the partitions they would form are small. Therefore, the cut associated with them would have a small cost, while the desired cut to properly segment an image may require cutting out a significant part of the image.

54 Normalized Cut was introduced in 2000 to solve this issue. The key idea is that the cost of the cut should be normalized based on the component size. That way, the few or isolated nodes won t be cheap to cut anymore, since after normalization with their small size, their relative cost increases. The idea is implemented easily by updating the cost of a feasible cut to this equation. A and B are the two sub-graphs the cut would yield. Cut(A,B) is the traditional cost of cut defined before (i.e., the sum of the weights of the edges being removed). V is the set of all nodes in the graph. assoc(x,a) is the sum of the weights of all edges in the graph that touch the sub-graph A. Therefore, if the sub-graph the cut produces is small, its cost will be proportionally higher.

55 This would be a low cost (2) unnormalized cut. But now after normalization, the cost would increase since there is only one node in the resulted sub-graph.

56 On the other hand, this cut (which is actually the desired cut) after normalization will be low-cost due to the subgraph sizes. This cut wouldn t be prefered over the undesirable cut in the previous slide by the unnormalized cost (cost 2 vs 3).

57 Now, let s see how we can implement solving normalized cut. G(V,E,W) defines a graph where V are the nodes, E are the edges, and W are the edge weights. Each pixel is one node in V. x is a vector that represents a feasible cut. Each element in x represents one node in V; if the node is included in the cut, the corresponding element would be +1, and -1 otherwise. D is a diagonal matrix. D_i_i (the elements on the main diagonal) represents the i_th node and its value is equal to the summation of the weights of all edges connected to the i_th node. Now, k is the sum of weights of all the edges in A over the total weight of the graph. b is therefore the proportion of the weights of A versus the weights of B (remember that A and B are disjoint and their union is the entire graph, i.e. V). y is the reformulation of x based on the proportion of weights in the subgraph the cut defines. The elements included in A will be +1 while the rest will be -b. The two equations in the bottom left are direct facts that can be derived from the definitions explained above.

58 Now it can be shown that finding the optimal cut (i.e. x) is equal to solving the optimization problem shown in the top right. The full derivations can be found in the normalized cut paper (referenced in the homework). To make sure the x satisfying this optimization equation is a feasible cut, the assumptions shown below it (y_td1-0, etc) should be satisfied. Therefore, the optimization should be solved subject to these assumptions. The paper also shows that this optimization problem can be solved through an eigen system. The eigenvector associated with the second smaller (absolute) eigenvalue is an approximate solution for x. However, solving the eigen system may yield an infeasible cut (i.e., may not satisfy the conditions necessarily). In those cases, a subsequent quantization is usually employed to snap the found solution to the closest feasible cut.

59 We can intuitively interpret the idea of graph cuts (or generally segmentation) by looking at it as a dynamical system. Let s say we have a large number of particles connected to each other by springs. The springs have different elasticities and form local regions. If we shake this system, the particles that are connected with strong springs will vibrate together while the ones with looser springs will have a more dissimilar vibration pattern. The vibration modes we would observe in this system would be equal to the segments, if each particle was a pixel and the spring strength was proportional to the similarity between pixels.

60 Normalized cut is a generic and flexible framework that can be adopted for many problems, including segmentation, with decent results. However, it requires high amount of storage and computation. It also has a bias towards partitioning the graph into near-equal-sized partition (due to size normalization).

61 Let s briefly look at another way of performing segmentation based on graphs.

62 This time, let s assume we want to segment an image into its foreground and background (2 clusters) and we can utilize a quick help from the user. The user can draw a few lines/markers on the image specifying a few regions which belong to the foreground or background (white and red strokes, respectively). How could we solve this problem now? How to incorporate the user input and encourage the segmentation results to be consistent with it and yield the foreground and background?

63 We can solve this by defining a labeling problem (L) in which we label each pixel with either 0 or 1, representing background and foreground, respectively. We then define an energy function that assigns an energy to each feasible labeling instance (L). The labeling instance L with the lowest energy corresponds to the desired segmentation. In high-level, there are two terms in this energy function, and they re mixed using a linear mixture. Lambda is the linear mixture constant specifying the contribution of each term in the final energy. The first terms is the match cost which specifies how likely it is for one pixel to be labeled as either foreground or background. This is usually acquired through matching the similarity of the pixel to the user specified regions. The second term is the smoothness cost. This is similar to the coherency cues we discussed in the previous segmentation methods. This term encourages the segments to be smooth and coherent.

64 For a given image with sample user inputs, here we show the corresponding energies for each pixel to belong to either foreground or background (i.e., the match cost in the energy function). The local inconsistencies and spurious outliers are expected to be recovered using the smoothness cost.

65 One popular example of this class of segmentation methods is GrabCut, introduced in 2004 by Rother et al.. It receives a boundary from the user for the foreground and segments it out. It works quite well and is implemented in Microsoft PowerPoint! The Remove Background feature in PowerPoint is in fact GrabCut.

66 What we discussed so far was almost entirely bottom up segmentation. As you saw, the resulting segments might not have any high level and semantic meaning. Let s look at a couple of semantic segmentation methods briefly. You ll see more about this topic and scene understanding in the next lectures.

67 Semantic segmentation seeks the segments in the image that have a semantic meaning, e.g., all pixels on one object, while the appearance of those pixels may not be that coherent. In order to solve this problem, we need to define semantics and develop an algorithm that has a notion of it. This is usually where machine learning comes into the picture. One recent and popular way of solving this problem is using fully convolutional neural networks. The neural network receives an image as its input and returns a perpixel mask in the output, where each pixel is labeled with its semantic class. This is solved through a series of consecutive convolution operations. The parameters of the convolutions are entirely learned using a fully supervised dataset; that s why they re expected to have a notion of semantics, objects, and whatever high-level label they ve seen in their training data.

68 Duygulu et al. also proposed a nominal method in this context back in The overall goal is similar to what we discussed so far in topdown segmentation. Their method is centered around finding a mapping between a set of regions in the image and keywords supplied with the image using EM. The result is the regions in the image along with their corresponding semantic keyword.

69 Another popular method for semantic segmentation belongs to Ladicky et al. While their overall goal is the same as previously overviewed methods, their main distinction is using a hierarchical random field model that enables integration of features from various levels of quantization from the image. There are a large number of other well-known semantic segmentation methods out there that we don t have time to overview in this lecture, but the few that we showed should establish what they seek and what their conceptual requirements are.

Lecture 16 Segmentation and Scene understanding

Lecture 16 Segmentation and Scene understanding Lecture 16 Segmentation and Scene understanding Introduction! Mean-shift! Graph-based segmentation! Top-down segmentation! Silvio Savarese Lecture 15 -! 3-Mar-14 Segmentation Silvio Savarese Lecture 15

More information

Applications. Foreground / background segmentation Finding skin-colored regions. Finding the moving objects. Intelligent scissors

Applications. Foreground / background segmentation Finding skin-colored regions. Finding the moving objects. Intelligent scissors Segmentation I Goal Separate image into coherent regions Berkeley segmentation database: http://www.eecs.berkeley.edu/research/projects/cs/vision/grouping/segbench/ Slide by L. Lazebnik Applications Intelligent

More information

Image Segmentation. Selim Aksoy. Bilkent University

Image Segmentation. Selim Aksoy. Bilkent University Image Segmentation Selim Aksoy Department of Computer Engineering Bilkent University saksoy@cs.bilkent.edu.tr Examples of grouping in vision [http://poseidon.csd.auth.gr/lab_research/latest/imgs/s peakdepvidindex_img2.jpg]

More information

CS 534: Computer Vision Segmentation and Perceptual Grouping

CS 534: Computer Vision Segmentation and Perceptual Grouping CS 534: Computer Vision Segmentation and Perceptual Grouping Ahmed Elgammal Dept of Computer Science CS 534 Segmentation - 1 Outlines Mid-level vision What is segmentation Perceptual Grouping Segmentation

More information

Image Segmentation. Selim Aksoy. Bilkent University

Image Segmentation. Selim Aksoy. Bilkent University Image Segmentation Selim Aksoy Department of Computer Engineering Bilkent University saksoy@cs.bilkent.edu.tr Examples of grouping in vision [http://poseidon.csd.auth.gr/lab_research/latest/imgs/s peakdepvidindex_img2.jpg]

More information

The goals of segmentation

The goals of segmentation Image segmentation The goals of segmentation Group together similar-looking pixels for efficiency of further processing Bottom-up process Unsupervised superpixels X. Ren and J. Malik. Learning a classification

More information

Segmentation & Clustering

Segmentation & Clustering EECS 442 Computer vision Segmentation & Clustering Segmentation in human vision K-mean clustering Mean-shift Graph-cut Reading: Chapters 14 [FP] Some slides of this lectures are courtesy of prof F. Li,

More information

Segmentation and Grouping

Segmentation and Grouping CS 1699: Intro to Computer Vision Segmentation and Grouping Prof. Adriana Kovashka University of Pittsburgh September 24, 2015 Goals: Grouping in vision Gather features that belong together Obtain an intermediate

More information

6.801/866. Segmentation and Line Fitting. T. Darrell

6.801/866. Segmentation and Line Fitting. T. Darrell 6.801/866 Segmentation and Line Fitting T. Darrell Segmentation and Line Fitting Gestalt grouping Background subtraction K-Means Graph cuts Hough transform Iterative fitting (Next time: Probabilistic segmentation)

More information

Segmentation (continued)

Segmentation (continued) Segmentation (continued) Lecture 05 Computer Vision Material Citations Dr George Stockman Professor Emeritus, Michigan State University Dr Mubarak Shah Professor, University of Central Florida The Robotics

More information

Segmentation and Grouping April 19 th, 2018

Segmentation and Grouping April 19 th, 2018 Segmentation and Grouping April 19 th, 2018 Yong Jae Lee UC Davis Features and filters Transforming and describing images; textures, edges 2 Grouping and fitting [fig from Shi et al] Clustering, segmentation,

More information

Outline. Segmentation & Grouping. Examples of grouping in vision. Grouping in vision. Grouping in vision 2/9/2011. CS 376 Lecture 7 Segmentation 1

Outline. Segmentation & Grouping. Examples of grouping in vision. Grouping in vision. Grouping in vision 2/9/2011. CS 376 Lecture 7 Segmentation 1 Outline What are grouping problems in vision? Segmentation & Grouping Wed, Feb 9 Prof. UT-Austin Inspiration from human perception Gestalt properties Bottom-up segmentation via clustering Algorithms: Mode

More information

Lecture 10: Semantic Segmentation and Clustering

Lecture 10: Semantic Segmentation and Clustering Lecture 10: Semantic Segmentation and Clustering Vineet Kosaraju, Davy Ragland, Adrien Truong, Effie Nehoran, Maneekwan Toyungyernsub Department of Computer Science Stanford University Stanford, CA 94305

More information

CS 534: Computer Vision Segmentation and Perceptual Grouping

CS 534: Computer Vision Segmentation and Perceptual Grouping CS 534: Computer Vision Segmentation and Perceptual Grouping Spring 2005 Ahmed Elgammal Dept of Computer Science CS 534 Segmentation - 1 Where are we? Image Formation Human vision Cameras Geometric Camera

More information

Grouping and Segmentation

Grouping and Segmentation Grouping and Segmentation CS 554 Computer Vision Pinar Duygulu Bilkent University (Source:Kristen Grauman ) Goals: Grouping in vision Gather features that belong together Obtain an intermediate representation

More information

Lecture: k-means & mean-shift clustering

Lecture: k-means & mean-shift clustering Lecture: k-means & mean-shift clustering Juan Carlos Niebles and Ranjay Krishna Stanford Vision and Learning Lab Lecture 11-1 Recap: Image Segmentation Goal: identify groups of pixels that go together

More information

CMPSCI 670: Computer Vision! Grouping

CMPSCI 670: Computer Vision! Grouping CMPSCI 670: Computer Vision! Grouping University of Massachusetts, Amherst October 14, 2014 Instructor: Subhransu Maji Slides credit: Kristen Grauman and others Final project guidelines posted Milestones

More information

Lecture: k-means & mean-shift clustering

Lecture: k-means & mean-shift clustering Lecture: k-means & mean-shift clustering Juan Carlos Niebles and Ranjay Krishna Stanford Vision and Learning Lab 1 Recap: Image Segmentation Goal: identify groups of pixels that go together 2 Recap: Gestalt

More information

Segmentation. Bottom Up Segmentation

Segmentation. Bottom Up Segmentation Segmentation Bottom up Segmentation Semantic Segmentation Bottom Up Segmentation 1 Segmentation as clustering Depending on what we choose as the feature space, we can group pixels in different ways. Grouping

More information

Announcements. Image Segmentation. From images to objects. Extracting objects. Status reports next Thursday ~5min presentations in class

Announcements. Image Segmentation. From images to objects. Extracting objects. Status reports next Thursday ~5min presentations in class Image Segmentation Announcements Status reports next Thursday ~5min presentations in class Project voting From Sandlot Science Today s Readings Forsyth & Ponce, Chapter 1 (plus lots of optional references

More information

CS 664 Segmentation. Daniel Huttenlocher

CS 664 Segmentation. Daniel Huttenlocher CS 664 Segmentation Daniel Huttenlocher Grouping Perceptual Organization Structural relationships between tokens Parallelism, symmetry, alignment Similarity of token properties Often strong psychophysical

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

Segmentation and Grouping

Segmentation and Grouping 02/23/10 Segmentation and Grouping Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem Last week Clustering EM Today s class More on EM Segmentation and grouping Gestalt cues By boundaries

More information

EE 701 ROBOT VISION. Segmentation

EE 701 ROBOT VISION. Segmentation EE 701 ROBOT VISION Regions and Image Segmentation Histogram-based Segmentation Automatic Thresholding K-means Clustering Spatial Coherence Merging and Splitting Graph Theoretic Segmentation Region Growing

More information

Grouping and Segmentation

Grouping and Segmentation 03/17/15 Grouping and Segmentation Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem Today s class Segmentation and grouping Gestalt cues By clustering (mean-shift) By boundaries (watershed)

More information

CS 2770: Computer Vision. Edges and Segments. Prof. Adriana Kovashka University of Pittsburgh February 21, 2017

CS 2770: Computer Vision. Edges and Segments. Prof. Adriana Kovashka University of Pittsburgh February 21, 2017 CS 2770: Computer Vision Edges and Segments Prof. Adriana Kovashka University of Pittsburgh February 21, 2017 Edges vs Segments Figure adapted from J. Hays Edges vs Segments Edges More low-level Don t

More information

human vision: grouping k-means clustering graph-theoretic clustering Hough transform line fitting RANSAC

human vision: grouping k-means clustering graph-theoretic clustering Hough transform line fitting RANSAC COS 429: COMPUTER VISON Segmentation human vision: grouping k-means clustering graph-theoretic clustering Hough transform line fitting RANSAC Reading: Chapters 14, 15 Some of the slides are credited to:

More information

Segmentation and Grouping April 21 st, 2015

Segmentation and Grouping April 21 st, 2015 Segmentation and Grouping April 21 st, 2015 Yong Jae Lee UC Davis Announcements PS0 grades are up on SmartSite Please put name on answer sheet 2 Features and filters Transforming and describing images;

More information

CS 4495 Computer Vision. Segmentation. Aaron Bobick (slides by Tucker Hermans) School of Interactive Computing. Segmentation

CS 4495 Computer Vision. Segmentation. Aaron Bobick (slides by Tucker Hermans) School of Interactive Computing. Segmentation CS 4495 Computer Vision Aaron Bobick (slides by Tucker Hermans) School of Interactive Computing Administrivia PS 4: Out but I was a bit late so due date pushed back to Oct 29. OpenCV now has real SIFT

More information

Image Segmentation continued Graph Based Methods. Some slides: courtesy of O. Capms, Penn State, J.Ponce and D. Fortsyth, Computer Vision Book

Image Segmentation continued Graph Based Methods. Some slides: courtesy of O. Capms, Penn State, J.Ponce and D. Fortsyth, Computer Vision Book Image Segmentation continued Graph Based Methods Some slides: courtesy of O. Capms, Penn State, J.Ponce and D. Fortsyth, Computer Vision Book Previously Binary segmentation Segmentation by thresholding

More information

What is Computer Vision?

What is Computer Vision? Perceptual Grouping in Computer Vision Gérard Medioni University of Southern California What is Computer Vision? Computer Vision Attempt to emulate Human Visual System Perceive visual stimuli with cameras

More information

CS 2750: Machine Learning. Clustering. Prof. Adriana Kovashka University of Pittsburgh January 17, 2017

CS 2750: Machine Learning. Clustering. Prof. Adriana Kovashka University of Pittsburgh January 17, 2017 CS 2750: Machine Learning Clustering Prof. Adriana Kovashka University of Pittsburgh January 17, 2017 What is clustering? Grouping items that belong together (i.e. have similar features) Unsupervised:

More information

Targil 12 : Image Segmentation. Image segmentation. Why do we need it? Image segmentation

Targil 12 : Image Segmentation. Image segmentation. Why do we need it? Image segmentation Targil : Image Segmentation Image segmentation Many slides from Steve Seitz Segment region of the image which: elongs to a single object. Looks uniform (gray levels, color ) Have the same attributes (texture

More information

Image Segmentation. Shengnan Wang

Image Segmentation. Shengnan Wang Image Segmentation Shengnan Wang shengnan@cs.wisc.edu Contents I. Introduction to Segmentation II. Mean Shift Theory 1. What is Mean Shift? 2. Density Estimation Methods 3. Deriving the Mean Shift 4. Mean

More information

Lecture 11: E-M and MeanShift. CAP 5415 Fall 2007

Lecture 11: E-M and MeanShift. CAP 5415 Fall 2007 Lecture 11: E-M and MeanShift CAP 5415 Fall 2007 Review on Segmentation by Clustering Each Pixel Data Vector Example (From Comanciu and Meer) Review of k-means Let's find three clusters in this data These

More information

2D image segmentation based on spatial coherence

2D image segmentation based on spatial coherence 2D image segmentation based on spatial coherence Václav Hlaváč Czech Technical University in Prague Center for Machine Perception (bridging groups of the) Czech Institute of Informatics, Robotics and Cybernetics

More information

EE795: Computer Vision and Intelligent Systems

EE795: Computer Vision and Intelligent Systems EE795: Computer Vision and Intelligent Systems Spring 2012 TTh 17:30-18:45 FDH 204 Lecture 14 130307 http://www.ee.unlv.edu/~b1morris/ecg795/ 2 Outline Review Stereo Dense Motion Estimation Translational

More information

Segmentation & Grouping Kristen Grauman UT Austin. Announcements

Segmentation & Grouping Kristen Grauman UT Austin. Announcements Segmentation & Grouping Kristen Grauman UT Austin Tues Feb 7 A0 on Canvas Announcements No office hours today TA office hours this week as usual Guest lecture Thursday by Suyog Jain Interactive segmentation

More information

Unsupervised learning in Vision

Unsupervised learning in Vision Chapter 7 Unsupervised learning in Vision The fields of Computer Vision and Machine Learning complement each other in a very natural way: the aim of the former is to extract useful information from visual

More information

Content-based Image and Video Retrieval. Image Segmentation

Content-based Image and Video Retrieval. Image Segmentation Content-based Image and Video Retrieval Vorlesung, SS 2011 Image Segmentation 2.5.2011 / 9.5.2011 Image Segmentation One of the key problem in computer vision Identification of homogenous region in the

More information

Lecture 7: Segmentation. Thursday, Sept 20

Lecture 7: Segmentation. Thursday, Sept 20 Lecture 7: Segmentation Thursday, Sept 20 Outline Why segmentation? Gestalt properties, fun illusions and/or revealing examples Clustering Hierarchical K-means Mean Shift Graph-theoretic Normalized cuts

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

Image Analysis - Lecture 5

Image Analysis - Lecture 5 Texture Segmentation Clustering Review Image Analysis - Lecture 5 Texture and Segmentation Magnus Oskarsson Lecture 5 Texture Segmentation Clustering Review Contents Texture Textons Filter Banks Gabor

More information

Design and Analysis of Algorithms Prof. Madhavan Mukund Chennai Mathematical Institute. Week 02 Module 06 Lecture - 14 Merge Sort: Analysis

Design and Analysis of Algorithms Prof. Madhavan Mukund Chennai Mathematical Institute. Week 02 Module 06 Lecture - 14 Merge Sort: Analysis Design and Analysis of Algorithms Prof. Madhavan Mukund Chennai Mathematical Institute Week 02 Module 06 Lecture - 14 Merge Sort: Analysis So, we have seen how to use a divide and conquer strategy, we

More information

SYDE Winter 2011 Introduction to Pattern Recognition. Clustering

SYDE Winter 2011 Introduction to Pattern Recognition. Clustering SYDE 372 - Winter 2011 Introduction to Pattern Recognition Clustering Alexander Wong Department of Systems Design Engineering University of Waterloo Outline 1 2 3 4 5 All the approaches we have learned

More information

Computer Vision 5 Segmentation by Clustering

Computer Vision 5 Segmentation by Clustering Computer Vision 5 Segmentation by Clustering MAP-I Doctoral Programme Miguel Tavares Coimbra Outline Introduction Applications Simple clustering K-means clustering Graph-theoretic clustering Acknowledgements:

More information

Shadows in the graphics pipeline

Shadows in the graphics pipeline Shadows in the graphics pipeline Steve Marschner Cornell University CS 569 Spring 2008, 19 February There are a number of visual cues that help let the viewer know about the 3D relationships between objects

More information

Digital Image Processing

Digital Image Processing Digital Image Processing Segmentation Material in this presentation is largely based on/derived from presentations by: Sventlana Lazebnik, and Noah Snavely Brent M. Dingle, Ph.D. 2015 Game Design and Development

More information

CS443: Digital Imaging and Multimedia Perceptual Grouping Detecting Lines and Simple Curves

CS443: Digital Imaging and Multimedia Perceptual Grouping Detecting Lines and Simple Curves CS443: Digital Imaging and Multimedia Perceptual Grouping Detecting Lines and Simple Curves Spring 2008 Ahmed Elgammal Dept. of Computer Science Rutgers University Outlines Perceptual Grouping and Segmentation

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

CS4670 / 5670: Computer Vision Noah Snavely

CS4670 / 5670: Computer Vision Noah Snavely { { 11/26/2013 CS4670 / 5670: Computer Vision Noah Snavely Graph-Based Image Segmentation Stereo as a minimization problem match cost Want each pixel to find a good match in the other image smoothness

More information

(Refer Slide Time: 0:51)

(Refer Slide Time: 0:51) Introduction to Remote Sensing Dr. Arun K Saraf Department of Earth Sciences Indian Institute of Technology Roorkee Lecture 16 Image Classification Techniques Hello everyone welcome to 16th lecture in

More information

CS 664 Slides #11 Image Segmentation. Prof. Dan Huttenlocher Fall 2003

CS 664 Slides #11 Image Segmentation. Prof. Dan Huttenlocher Fall 2003 CS 664 Slides #11 Image Segmentation Prof. Dan Huttenlocher Fall 2003 Image Segmentation Find regions of image that are coherent Dual of edge detection Regions vs. boundaries Related to clustering problems

More information

From Pixels to Blobs

From Pixels to Blobs From Pixels to Blobs 15-463: Rendering and Image Processing Alexei Efros Today Blobs Need for blobs Extracting blobs Image Segmentation Working with binary images Mathematical Morphology Blob properties

More information

IMAGE SEGMENTATION. Václav Hlaváč

IMAGE SEGMENTATION. Václav Hlaváč IMAGE SEGMENTATION Václav Hlaváč Czech Technical University in Prague Faculty of Electrical Engineering, Department of Cybernetics Center for Machine Perception http://cmp.felk.cvut.cz/ hlavac, hlavac@fel.cvut.cz

More information

Classification. Vladimir Curic. Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University

Classification. Vladimir Curic. Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University Classification Vladimir Curic Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University Outline An overview on classification Basics of classification How to choose appropriate

More information

CS 1675 Introduction to Machine Learning Lecture 18. Clustering. Clustering. Groups together similar instances in the data sample

CS 1675 Introduction to Machine Learning Lecture 18. Clustering. Clustering. Groups together similar instances in the data sample CS 1675 Introduction to Machine Learning Lecture 18 Clustering Milos Hauskrecht milos@cs.pitt.edu 539 Sennott Square Clustering Groups together similar instances in the data sample Basic clustering problem:

More information

Using the Kolmogorov-Smirnov Test for Image Segmentation

Using the Kolmogorov-Smirnov Test for Image Segmentation Using the Kolmogorov-Smirnov Test for Image Segmentation Yong Jae Lee CS395T Computational Statistics Final Project Report May 6th, 2009 I. INTRODUCTION Image segmentation is a fundamental task in computer

More information

CS 556: Computer Vision. Lecture 18

CS 556: Computer Vision. Lecture 18 CS 556: Computer Vision Lecture 18 Prof. Sinisa Todorovic sinisa@eecs.oregonstate.edu 1 Color 2 Perception of Color The sensation of color is caused by the brain Strongly affected by: Other nearby colors

More information

Project and Production Management Prof. Arun Kanda Department of Mechanical Engineering Indian Institute of Technology, Delhi

Project and Production Management Prof. Arun Kanda Department of Mechanical Engineering Indian Institute of Technology, Delhi Project and Production Management Prof. Arun Kanda Department of Mechanical Engineering Indian Institute of Technology, Delhi Lecture - 8 Consistency and Redundancy in Project networks In today s lecture

More information

Normalized cuts and image segmentation

Normalized cuts and image segmentation Normalized cuts and image segmentation Department of EE University of Washington Yeping Su Xiaodan Song Normalized Cuts and Image Segmentation, IEEE Trans. PAMI, August 2000 5/20/2003 1 Outline 1. Image

More information

Clustering Part 4 DBSCAN

Clustering Part 4 DBSCAN Clustering Part 4 Dr. Sanjay Ranka Professor Computer and Information Science and Engineering University of Florida, Gainesville DBSCAN DBSCAN is a density based clustering algorithm Density = number of

More information

Image Analysis Lecture Segmentation. Idar Dyrdal

Image Analysis Lecture Segmentation. Idar Dyrdal Image Analysis Lecture 9.1 - Segmentation Idar Dyrdal Segmentation Image segmentation is the process of partitioning a digital image into multiple parts The goal is to divide the image into meaningful

More information

Centrality Book. cohesion.

Centrality Book. cohesion. Cohesion The graph-theoretic terms discussed in the previous chapter have very specific and concrete meanings which are highly shared across the field of graph theory and other fields like social network

More information

Clustering CS 550: Machine Learning

Clustering CS 550: Machine Learning Clustering CS 550: Machine Learning This slide set mainly uses the slides given in the following links: http://www-users.cs.umn.edu/~kumar/dmbook/ch8.pdf http://www-users.cs.umn.edu/~kumar/dmbook/dmslides/chap8_basic_cluster_analysis.pdf

More information

Unsupervised Learning : Clustering

Unsupervised Learning : Clustering Unsupervised Learning : Clustering Things to be Addressed Traditional Learning Models. Cluster Analysis K-means Clustering Algorithm Drawbacks of traditional clustering algorithms. Clustering as a complex

More information

Segmentation by Clustering. Segmentation by Clustering Reading: Chapter 14 (skip 14.5) General ideas

Segmentation by Clustering. Segmentation by Clustering Reading: Chapter 14 (skip 14.5) General ideas Reading: Chapter 14 (skip 14.5) Data reduction - obtain a compact representation for interesting image data in terms of a set of components Find components that belong together (form clusters) Frame differencing

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

Segmentation by Clustering Reading: Chapter 14 (skip 14.5)

Segmentation by Clustering Reading: Chapter 14 (skip 14.5) Segmentation by Clustering Reading: Chapter 14 (skip 14.5) Data reduction - obtain a compact representation for interesting image data in terms of a set of components Find components that belong together

More information

Digital Image Processing. Prof. P.K. Biswas. Department of Electronics & Electrical Communication Engineering

Digital Image Processing. Prof. P.K. Biswas. Department of Electronics & Electrical Communication Engineering Digital Image Processing Prof. P.K. Biswas Department of Electronics & Electrical Communication Engineering Indian Institute of Technology, Kharagpur Image Segmentation - III Lecture - 31 Hello, welcome

More information

(Refer Slide Time: 00:02:00)

(Refer Slide Time: 00:02:00) Computer Graphics Prof. Sukhendu Das Dept. of Computer Science and Engineering Indian Institute of Technology, Madras Lecture - 18 Polyfill - Scan Conversion of a Polygon Today we will discuss the concepts

More information

University of Florida CISE department Gator Engineering. Clustering Part 4

University of Florida CISE department Gator Engineering. Clustering Part 4 Clustering Part 4 Dr. Sanjay Ranka Professor Computer and Information Science and Engineering University of Florida, Gainesville DBSCAN DBSCAN is a density based clustering algorithm Density = number of

More information

Image Segmentation continued Graph Based Methods

Image Segmentation continued Graph Based Methods Image Segmentation continued Graph Based Methods Previously Images as graphs Fully-connected graph node (vertex) for every pixel link between every pair of pixels, p,q affinity weight w pq for each link

More information

Face Detection. Gary Chern, Paul Gurney, and Jared Starman

Face Detection. Gary Chern, Paul Gurney, and Jared Starman Face Detection Gary Chern, Paul Gurney, and Jared Starman. Introduction Automatic face detection is a complex problem in image processing. Many methods exist to solve this problem such as template matching,

More information

Lecture 3: Linear Classification

Lecture 3: Linear Classification Lecture 3: Linear Classification Roger Grosse 1 Introduction Last week, we saw an example of a learning task called regression. There, the goal was to predict a scalar-valued target from a set of features.

More information

9.913 Pattern Recognition for Vision. Class 8-2 An Application of Clustering. Bernd Heisele

9.913 Pattern Recognition for Vision. Class 8-2 An Application of Clustering. Bernd Heisele 9.913 Class 8-2 An Application of Clustering Bernd Heisele Fall 2003 Overview Problem Background Clustering for Tracking Examples Literature Homework Problem Detect objects on the road: Cars, trucks, motorbikes,

More information

Lab 9. Julia Janicki. Introduction

Lab 9. Julia Janicki. Introduction Lab 9 Julia Janicki Introduction My goal for this project is to map a general land cover in the area of Alexandria in Egypt using supervised classification, specifically the Maximum Likelihood and Support

More information

CPSC 425: Computer Vision

CPSC 425: Computer Vision 1 / 31 CPSC 425: Computer Vision Instructor: Jim Little little@cs.ubc.ca Department of Computer Science University of British Columbia Lecture Notes 2016/2017 Term 2 2 / 31 Menu March 16, 2017 Topics:

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

Case-Based Reasoning. CS 188: Artificial Intelligence Fall Nearest-Neighbor Classification. Parametric / Non-parametric.

Case-Based Reasoning. CS 188: Artificial Intelligence Fall Nearest-Neighbor Classification. Parametric / Non-parametric. CS 188: Artificial Intelligence Fall 2008 Lecture 25: Kernels and Clustering 12/2/2008 Dan Klein UC Berkeley Case-Based Reasoning Similarity for classification Case-based reasoning Predict an instance

More information

CS 188: Artificial Intelligence Fall 2008

CS 188: Artificial Intelligence Fall 2008 CS 188: Artificial Intelligence Fall 2008 Lecture 25: Kernels and Clustering 12/2/2008 Dan Klein UC Berkeley 1 1 Case-Based Reasoning Similarity for classification Case-based reasoning Predict an instance

More information

Data Structures and Algorithms Dr. Naveen Garg Department of Computer Science and Engineering Indian Institute of Technology, Delhi.

Data Structures and Algorithms Dr. Naveen Garg Department of Computer Science and Engineering Indian Institute of Technology, Delhi. Data Structures and Algorithms Dr. Naveen Garg Department of Computer Science and Engineering Indian Institute of Technology, Delhi Lecture 18 Tries Today we are going to be talking about another data

More information

Lecture 6: Edge Detection

Lecture 6: Edge Detection #1 Lecture 6: Edge Detection Saad J Bedros sbedros@umn.edu Review From Last Lecture Options for Image Representation Introduced the concept of different representation or transformation Fourier Transform

More information

Clustering. Subhransu Maji. CMPSCI 689: Machine Learning. 2 April April 2015

Clustering. Subhransu Maji. CMPSCI 689: Machine Learning. 2 April April 2015 Clustering Subhransu Maji CMPSCI 689: Machine Learning 2 April 2015 7 April 2015 So far in the course Supervised learning: learning with a teacher You had training data which was (feature, label) pairs

More information

Fundamentals of Stereo Vision Michael Bleyer LVA Stereo Vision

Fundamentals of Stereo Vision Michael Bleyer LVA Stereo Vision Fundamentals of Stereo Vision Michael Bleyer LVA Stereo Vision What Happened Last Time? Human 3D perception (3D cinema) Computational stereo Intuitive explanation of what is meant by disparity Stereo matching

More information

Computing Visibility. Backface Culling for General Visibility. One More Trick with Planes. BSP Trees Ray Casting Depth Buffering Quiz

Computing Visibility. Backface Culling for General Visibility. One More Trick with Planes. BSP Trees Ray Casting Depth Buffering Quiz Computing Visibility BSP Trees Ray Casting Depth Buffering Quiz Power of Plane Equations We ve gotten a lot of mileage out of one simple equation. Basis for D outcode-clipping Basis for plane-at-a-time

More information

Soft shadows. Steve Marschner Cornell University CS 569 Spring 2008, 21 February

Soft shadows. Steve Marschner Cornell University CS 569 Spring 2008, 21 February Soft shadows Steve Marschner Cornell University CS 569 Spring 2008, 21 February Soft shadows are what we normally see in the real world. If you are near a bare halogen bulb, a stage spotlight, or other

More information

Prof. Fanny Ficuciello Robotics for Bioengineering Visual Servoing

Prof. Fanny Ficuciello Robotics for Bioengineering Visual Servoing Visual servoing vision allows a robotic system to obtain geometrical and qualitative information on the surrounding environment high level control motion planning (look-and-move visual grasping) low level

More information

CPSC 340: Machine Learning and Data Mining. Density-Based Clustering Fall 2016

CPSC 340: Machine Learning and Data Mining. Density-Based Clustering Fall 2016 CPSC 340: Machine Learning and Data Mining Density-Based Clustering Fall 2016 Assignment 1 : Admin 2 late days to hand it in before Wednesday s class. 3 late days to hand it in before Friday s class. 0

More information

Clustering. So far in the course. Clustering. Clustering. Subhransu Maji. CMPSCI 689: Machine Learning. dist(x, y) = x y 2 2

Clustering. So far in the course. Clustering. Clustering. Subhransu Maji. CMPSCI 689: Machine Learning. dist(x, y) = x y 2 2 So far in the course Clustering Subhransu Maji : Machine Learning 2 April 2015 7 April 2015 Supervised learning: learning with a teacher You had training data which was (feature, label) pairs and the goal

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

Region-based Segmentation

Region-based Segmentation Region-based Segmentation Image Segmentation Group similar components (such as, pixels in an image, image frames in a video) to obtain a compact representation. Applications: Finding tumors, veins, etc.

More information

Feature Extractors. CS 188: Artificial Intelligence Fall Nearest-Neighbor Classification. The Perceptron Update Rule.

Feature Extractors. CS 188: Artificial Intelligence Fall Nearest-Neighbor Classification. The Perceptron Update Rule. CS 188: Artificial Intelligence Fall 2007 Lecture 26: Kernels 11/29/2007 Dan Klein UC Berkeley Feature Extractors A feature extractor maps inputs to feature vectors Dear Sir. First, I must solicit your

More information

HOUGH TRANSFORM CS 6350 C V

HOUGH TRANSFORM CS 6350 C V HOUGH TRANSFORM CS 6350 C V HOUGH TRANSFORM The problem: Given a set of points in 2-D, find if a sub-set of these points, fall on a LINE. Hough Transform One powerful global method for detecting edges

More information

Finally: Motion and tracking. Motion 4/20/2011. CS 376 Lecture 24 Motion 1. Video. Uses of motion. Motion parallax. Motion field

Finally: Motion and tracking. Motion 4/20/2011. CS 376 Lecture 24 Motion 1. Video. Uses of motion. Motion parallax. Motion field Finally: Motion and tracking Tracking objects, video analysis, low level motion Motion Wed, April 20 Kristen Grauman UT-Austin Many slides adapted from S. Seitz, R. Szeliski, M. Pollefeys, and S. Lazebnik

More information

CS1114 Section 8: The Fourier Transform March 13th, 2013

CS1114 Section 8: The Fourier Transform March 13th, 2013 CS1114 Section 8: The Fourier Transform March 13th, 2013 http://xkcd.com/26 Today you will learn about an extremely useful tool in image processing called the Fourier transform, and along the way get more

More information

Tracking Computer Vision Spring 2018, Lecture 24

Tracking Computer Vision Spring 2018, Lecture 24 Tracking http://www.cs.cmu.edu/~16385/ 16-385 Computer Vision Spring 2018, Lecture 24 Course announcements Homework 6 has been posted and is due on April 20 th. - Any questions about the homework? - How

More information

CS4670: Computer Vision

CS4670: Computer Vision CS4670: Computer Vision Noah Snavely Lecture 34: Segmentation From Sandlot Science Announcements In-class exam this Friday, December 3 Review session in class on Wednesday Final projects: Slides due: Sunday,

More information

Dr. Ulas Bagci

Dr. Ulas Bagci CAP5415-Computer Vision Lecture 11-Image Segmentation (BASICS): Thresholding, Region Growing, Clustering Dr. Ulas Bagci bagci@ucf.edu 1 Image Segmentation Aim: to partition an image into a collection of

More information

Programming assignment 3 Mean-shift

Programming assignment 3 Mean-shift Programming assignment 3 Mean-shift 1 Basic Implementation The Mean Shift algorithm clusters a d-dimensional data set by associating each point to a peak of the data set s probability density function.

More information