1. For more than a decade peoplehave developed an interest forPROJECT the researching in the domainSEMINAR/GROUPNO.-7 IMAGE PROCESSING and now theres more than a millionpatents have been submittedthis year on this topic.Andnow we bring youOUR PROJECT ONIMAGEPROCESSING

2. PART One1INTRO TOIMAGEprocessing 3. Introduction Image processing is any form of signal processingfor which the input is an image, such asphotographs or frames of video; the output ofimage processing can be either an image or a set ofcharacteristics or parameters related to the image. Most image-processing techniques involve treatingthe image as a two-dimensional signal andapplying standard signal-processing techniques to it. 4. Types Image processing usually refers to digital image processing, but optical and analog image processing are also possible. 5. Image Processing Operations Geometric transformations such as enlargement,reduction, and rotation Color corrections such as brightness and contrastadjustments, quantization, or conversion to a differentcolor space Digital compositing or optical compositing (combinationof two or more images). Used in filmmaking to make a"matte" Interpolation, demosaicing, and recovery of a full imagefrom a raw image format using a Bayer filter pattern 6. Image ProcessingOperations(Contd.) Image editing (e.g., to increase the quality of a digitalimage) Image differencing Image registration (alignment of two or more images) Image stabilization Extending dynamic range by combining differently exposedimages 7. Image Processing Applications Computer vision Face recognition Feature detection Non-photorealistic rendering Medical image processing Microscope image processing Morphological image processing Remote sensing 8. IMAGE SEGMENTATION USING WAVELET TRANSFORMATION& FUZZY LOGIC 9. 2 PART IMAGE 10. The Definition The purpose of image segmentation is to partition animage into meaningful regions with respect to a particularapplication . Image segmentation is the process of dividing animage into different regions such that each region ishomogeneous. It basically identifies the pixels belonging to thedesired object that we may want to extract from aninput image. 11. More about segmentation To humans, an image is not just a randomcollection of pixels; it is a meaningfularrangement of regions and objects. There also exits a variety of images: naturalscenes, paintings, etc. Despite the largevariations of these images, humans have noproblem to interpret them. 12. Segmentation can beIN TERMS OF:Grey level ColourTexture, Depth or motion 13. Introduction to imagesegmentation Example 1 Segmentation based on greyscale Very simple model of greyscale leads to inaccuracies in object labelling 13 14. Introduction to imagesegmentation Example 2 Segmentation based on texture Enables object surfaces with varying patterns of greyto be segmented 14 15. Introduction to imagesegmentation15 16. Introduction to imagesegmentation Example 3 Segmentation based on motion The main difficulty of motion segmentation is that an intermediate step is required to (either implicitly or explicitly) estimate an optical flow field The segmentation must be based on this estimate and not, in general, the true flow16 17. Introduction to imagesegmentation17 18. TYPES OF IMAGE SEGMENTATIONA) Supervised.- These methods require the interactivity inwhich the pixels belonging to the same intensity rangepointed out manually and segmented.B) Automatic.- This is also known as unsupervised methods,where the algorithms need some prior information, so thesemethods are more complex.C) Semi-automatic.- That is the combination of manual andautomatic segmentation. 19. SEGMENTING CAN ALSO BE ONDISCONTINUITY : Partitioning an image basedon abrupt change.Edge detection in a image.SIMILARITY : Partitioning an image intoregions that are similar according to a set of predefined criteria.Thresholding , Region Growing , Clustering. 20. DISCONTINUITIESLINESPOINTSEDGES 21. Definition of edge Definition : Set of connected pixels that lie on theboundary b/w 2 regions. Edge is a local concept & boundary is globalconcept. Reasonable definition of edge requires ability tomeasure gray level transition in a meaningful way. 22. EDGE DETECTION It is the most common approach fordetecting meaningful discontinuities ingray level. Process: By implementing the 1st orderderivative 2nd order derivative ,edges inan image can be detected. 23. DIFFERENCE B/W EDGE & PRACTICAL EDGE IDEAL : Set of pixels ,each of which is located at anorthogonal step transition in gray level PRACTICAL : Used by optics sampling and other imageacquisition imperfections and yield blurred edgeswhere degree of blurring is determined by factors suchas1. Quality of image acquisition system2. Sampling rate3. Illumination conditions under which image is acquired 24. EXAMPLE 25. Segmentation Techniques There are 2 very simple image segmentation techniques that are based on the grey level histogram of an image Thresholding Clustering But in our project we will be using clustering so we willlook into the details of clustering. 26. Clustering. Similar data pointsgrouped together intoclusters. In this , centroid is usedto represent each cluster,based on the similaritywith the centroid ofcluster we can classifythe patterns. 27. ClusteringMost popular clustering algorithms suffer from two majordrawbacks First, the number of clusters is predefined, which makesthem inadequate for batch processing of huge imagedatabases Secondly, the clusters are represented by their centroid andbuilt using an Euclidean distance therefore inducinggenerally an hyperspheric cluster shape, which makes themunable to capture the real structure of the data. This is especially true in the case of color clustering whereclusters are arbitrarily shaped 28. Clustering Algorithms K-means K-medoids Hierarchical Clustering There are many other algorithms used for clustering.Here we would look into 2 algorithms mainly K-meansAnd Hierarchical Clustering. 29. HIERARCHICAL CLUSTERING The concept of hierarchical clustering is to construct adendrogram representing the nested grouping ofpatterns (for image, known as pixels) and thesimilarity levels at which groupings change. We can apply the two-dimensional data set to interpretthe operation of the hierarchical clustering algorithm 30. CLUSTER & DENDOGRAM 31. K-means Clustering Algorithm Step1. Determine the number of clusters we want in the finalclassified result and set the number as N. Randomly select Npatterns in the whole data bases as the N centroids of N clusters. Step2. Classify each pattern to the closest cluster centroid. Theclosest usually represent the pixel value is similarity, but it stillcan consider other features. Step3. Recompute the cluster centroids and then there have Ncentroids of N clusters as we do after Step1. Step4. Repeat the iteration of Step 2 to 3 until a convergencecriterion is met. 32. APPLICATIONS OF IMAGESEGEMENTATION Medical Imaging Tasks (detecting tumors,etc) Object recognitions in images of remote sensing viasatellite on aerial platforms. Automated recognition systems to inspect theelectronic assemblies Biometrics Automated traffic control system. 33. MODULE3WAVELET TRANS 34. More Facts about Wavelets : Wavelets are localized in frequency as well as in space having an advantageover the Fourier transform which is only localized in frequency As a result temporal-spatial information is maintained during the wavelettransformation process which is extremely important for edge detection.Two methods based on wavelets from the multiresolution point of view havebeen introduced - The first method was based on the two-dimensional fast wavelet transformusing the Biorthogonal Mother Wavelet The second method was based on a new wavelet named Contourletwhich has been developed recently as an improvement of the classicalwavelets. 35. Some facts about Fourier To Wavelet Analysis The Fourier transform has been the mainstay of transform-based imageprocessing since the late 1950s but they have a serious drawback as onlyfrequency information remains while the local one is lost which means changein Fourier coefficients has a global effect on the image. This means, that any modification of the Fourier coefficients has a global effecton the image. In order to involve localization on the analysis, the Short TimeFourier transform (STFT) is adapted. In this case, the image is windowed, and thus the information has a precisionrelevant to the size of the window used.The drawback is that the window is the same in all frequencies. Wavelet analysis allows the variation of the window based on the frequencyinformation. As a result, long time intervals are used in low-frequencyinformation and short time intervals in high-frequency information. 36. Fast Wavelet Transform using DWT : 37. 4 MODULE:FUZZY LOGIC 38. Some Fuzzy Background Fuzzy logic is an approach to computing based on"degrees of truth" rather than the usual "true or false" (1or 0) Boolean logic on which the modern computer isbased. The idea of fuzzy logic was first advanced by Dr.Lotfi Zadeh of the University of California at Berkeley inthe 1960s. Fuzzy logic includes 0 and 1 as extreme cases of truth (or"the state of matters" or "fact") but also includes thevarious states of truth in between 39. Fuzzy Vs. ProbabilityThe difference between probability and fuzzy logic is clearwhen we consider the underlying concept that each attemptsto model. Probability is concerned with the undecidability inthe outcome of clearly defined and randomly occurringevents, while fuzzy logic is concerned with the ambiguity orundecidability inherent in the description of the event itself.Fuzziness is often expressed as ambiguity rather thanimprecision or uncertainty and remains a characteristic ofperception as well as concept. 40. Membership Functions (MFs) What is a MF? Linguistic Variable A Normal MF attains 1 and 0 for some inputx1 , x2A x1 1, A x2 0 How do we construct MFs? Heuristic Rank ordering Mathematical Models Adaptive (Neural Networks, Genetic Algorithms ) 41. Membership Function Examples Gaussian x c 2 Sigmoid 2 2f gmf x; , c e1f smfx, a , c a x c1 eTriangularTrapezoidalx a c xx a d xf x; a, b, cmax min,,0 f x; a, b, c, dmax min ,1,,0b a c bb a d c 42. Example: Finding an ImageThreshold Membership Value 1 f smf x, a, c a x c 1 e Gray Level 43. Crisp Vs. Fuzzy Fuzzy Sets Crisp Sets Membership values on [0,1] True/False {0,1} Law of Excluded Middle and Non- Law of Excluded Middle and Non-Contradiction do not necessarily Contradiction hold:hold:AA AAAA AA Fuzzy Membership Function Crisp Membership Function Flexibility in choosing the Intersection (AND) , Union (OR),Intersection (T-Norm), Union (S- and Negation (NOT) are fixedNorm) and Negation operations 44. Image ProcessingBinaryGray LevelColor (RGB,HSV etc.) Can we give a crisp definition to light blue? 45. Feature Vector Feature Feature is any distinctive aspect, quality or characteristic Features may be symbolic (i.e., color) or numeric (i.e., height) The combination of d features is represented as a d-dimensionalcolumn vector called a feature vector The d-dimensional space defined by the feature vector is calledfeature space Objects are represented as points in feature space. Thisrepresentation is called a scatter plot 46. Fuzzy C-means ClusteringIn fuzzy clustering, each point has a degree of belonging to clusters, as in fuzzylogic, rather than belonging completely to just one cluster. Thus, points on theedge of a cluster, may be in the cluster to a lesser degree than points in the centerof cluster.Any point x has a set of coefficients giving the degree of being in the kth clusterwk(x). With fuzzy c-means, the centroid of a cluster is the mean of allpoints, weighted by their degree of belonging to the cluster: 47. Example: Finding Edges 21 ij mn min 1 , min ij , W i j1 ijg ijmax gijmin gij minij spatialspatial spatial ij ij 1 max gijmax gijmaxij spatial global spatial 48. Summary Fuzzy Logic can be useful in solving Human related tasks Evidence Theory gives tools to handle knowledge Membership functions and Aggregation methods can beselected according to the problem at hand Fuzzy logic can model nonlinear functions of arbitrarycomplexity. Fuzzy logic is tolerant of imprecise data. 49. MODULE5 50. AcknowledgementWe are thankful to our mentor Mr. Soumyadip Dhar for guidingus through our project . 51. The presentation wasbrought to you by Jishnu Mukherjee Lahaul Seth Rahul Kar