Please enable JavaScript.

Coggle requires JavaScript to display documents.

Computer Vision (Image Processing (Feature Matching (Point Correspondences…

- - - - Gaussian filter
      - Prewitt and Sobel filter
      - Box filter
  - - - Canny Edge Detection
        
        Non-maximum suppression - thin wide ridges down to single pixel width
        
        Hysterisis thresholding - define low & high thresholds, high to start edge curves, low to continue them
        
        Filter image with Derivative Of Gaussian
        
        Find magnitude and orientation of gradient
    - - Scale Invariant Feature Transform (SIFT)
        
        Filter
        
        Autocorrelation
        
        Threshold
        
        Select
        
        Difference of Gaussian
        
        Compute
        
        Histogram of gradient orientations
        
        Measurements relative to dominant orientation
        
        Descriptor
        
        4x4 array of 8 bin histograms
        128 negative numbers for each descriptor
  - - - RANSAC
        
        Estimate the model for the random subset
        
        Count the number of inliers that are within x of their predicted location
        
        Repeat the random selection process and keep the sample with the largest number of inliers
        
        Select a random subset of k correspondences
    - - Panorama Creation
    - - Object recognition
  - - - Disadvantages
        
        Intra-class variability
        
        Inter-class confusion
      - Algorithm
        
        Viloa & Jones Detection Algorithm
        
        Extracting haar like features
        
        2.Training model
        
        Classify
      - Rectangular HAAR like features provide a description of the window features of the image
      - Integral images allow fast computation of rectangle features
      - AdaBoost is used for feature selection and classification
      - Cascade of classifiers
        
        The classifiers become progressively more complex and have lower FP rates
        
        allows to obtain low FN rate in real-time
    - - Eigen Faces
        (compression-based feature-extraction)
        
        Principal Component Analysis is used to learn a set of basis component faces based on degrees of variance
        
        Normalization
        
        A face can be treated as a combination of component faces
        
        Eigenfaces can be differently weighted to represent any face
        
        Recognition
        
        Project unknown normalized face image into face space
        
        Compute the Euclidean distance D between the projection and all the projected training faces
        
        Closest face in the feature space considered a match or use KNN to predict class membership based on k closest samples
      - Linear Discriminant Analysis (LDA)
        
        LDA starts with criteria based on discrimination from the beginning
        
        LDA is a linear supervised feature extraction technique
        
        LDA seeks a new data representation, where points of the same class are as close as possible, while points of different classes are as far as possible
        
        LDA outperforms PCA when training set is large. Often use PCA then LDA.
        
        FisherFaces
        
        LDA Dimensionality reduction
- - - - Texture is a property of regions
      - Texture is a contextual property
      - Texture can be perceived at different scales or levels of resolution
      - A region presents a texture when the number of primitive patterns is large
    - - Use filters that look like patterns
        
        Filter Bank
        
        Collection of gaussian and linear filters
      - Consider magnitude of response
        
        Using mean absolute response
    - - Mean std deviation, histogram
- - - - Gestalt factors
        
        Factors that determine whether elements should be grouped together
        
        Familiar configuration
        
        Subjective contours
        
        Occlusion
      - Cluster similar pixels together
        
        Construct a vector of color and spatial coordinates
      - Clustering Techniques
        
        Agglomerative
        
        K-Means
        
        Mean-shift
        
        Algorithm
        
        Initialize windows at feature points
        
        Perform mean shift for each window until convergence
        
        Merge windows that end up at the same peak or mode
        
        Find features
        
        Advantages
        
        Does not assume spherical clusters
        
        Just a single parameter
        
        Finds variable number of modes
        
        Robust to outliers
        
        Generic technique
        
        Disadvantages
        
        Output depends on window size
        
        Computationally expensive
  - - - Given an initial contour near desired object, evolve the contour to fit the exact object boundary, like an Elastic band around a solid
        
        Elastic band is iteratively adjusted so as to:
        
        be near image positions with high gradients
        
        satisfy shape preferences of contour priors
      - Energy functions
        
        Internal
        
        Energy to encourage prior shape preferences.
        Energy pushing "out" on elastic band
        
        Tension, Elasticity
        
        Balloon force
        
        Stiffness, Curvature
        
        Shape Prior
        
        Externel
        
        Energy to encourage contour to fit on places where structures exist, Gradient-Based
        
        Energy "tightening" elastic band
        
        Idea is to minimize total energy (Einternal + Eextermal)
      - Disadvantages
        
        Depends on number and spacing of control points
        
        Snake may over-smooth the boundary
        
        The curve may intersect itself
        
        Can not follow topological changes of objects (e.g. splitting, merging)
        
        May get stuck in local minimum
        
        Parameters of energy function must be set well based on prior information
      - Advantages
        
        Good for non-rigid deformable objects: lips, hands, e.t.c
        
        Useful to track and fit non-rigid shapes
        
        Contour remains connected
        
        Possible to fill in subjective contours
        
        Flexible energy function definition
    - - Key Idea
        
        Evolve the curve outwards with a speed depending of the curvature and the image
        
        Quickly expand when passing over places with small image gradient
        
        Slow down when crossing large image gradient places
      - Comparison to Snakes
        
        Curve may change topology and form sharp corners
        
        Intrinsic geometric properties are easily determined
        
        2D and 3D formulation are the same
      - Initial contour must be completely contained within object
      - Allows topologically invariant segmentation
      - Independent on the initialization
    - - Function that captures how well reconstructed a function matches the original image
    - - Segments the image using model of interior and exterior
      - Heaviside step function is approximately 1 inside the contour, and 0 outside
      - Level set function is used to represent the curve
      - Assume interior and exterior both have approximately constant intensity
- - - - Camera Break, abrupt change between neighbouring frames
      - Gradual Transition over a set of consecutive frames
    - - Hard Cut
      - Fade
      - Dissolve
      - Wipe
    - - Find shots that differ and declare as shot-break
      - Cluster similar frames to identify as shots
      - Pixel comparison
        
        Simple, easy to implement
        
        Computationally Heavy
        
        Sensitive to camera motion
      - Block-based approach
        
        Performs better than pixel approach
        
        Cant identify dissolve, fade or fast moving objects
      - Histogram comparison
        
        More accurate than pixel-wise approach
        
        Compare number of occurrences for each color between subsequent frames
        
        Detects hard-cut, fade, wipe and dissolve
        
        Fails if two successive shots have the same histogram
        
        Cant distinguish fast object or camera motion
      - Edge change ratio
        
        No color information needed
        
        Compare number of entering edge and exiting edge pixels
        
        Better than other methods at detecting fade, wipe, dissolve and hard-cut
        
        Computationally heavy
        
        Fails when there is large amount of motion
      - Thresholding
        
        Calculate a time series function of discontinuity feature values for each frame.
        Pick cuts positions from the discontinuities function based on some thresholds
        
        Global Threshold
        
        A hard cut is declared every time the discontinuity value surpasses a global threshold
        
        Adaptive Threshold
        
        A soft cut is detected based on the difference of the current discontinuity values from its local neighbourhood
        
        Hardcut declared when discontinuity value is maximum in neighbourhood
  - - - Moving average to estimate background image (using the Median)
      - Subtract estimate from current frame
      - Large absolute values are interesting pixels, defined by a threshold
      - Use morphological operations to clean up pixels
    - - Apperture problem
        
        Pixels in a flat image region may appear stationary even if they are moving
      - Occlusion problem
        
        Erroneous labels may be assigned to pixels in covered or uncovered regions
    - - Simple & Fast
    - - Needs many images
      - Needs static background
      - Noisy segmentation
- - - - Quantize features
        
        Use a distance based classifier (KNN / SVM) to classify new images
        
        Using a spatial histogram
        
        Apply a SVM classifier
        
        Effective in high dimensional spaces
        
        Can use a custom kernel map
        
        Linear / additive / additive RBG kernels
        
        Computation time
        
        No direct multi-class svm
      - Learn Vocabulary (Train)
        
        Train classifier by counting occurence of features in labelled images
        
        Clustering the feature descriptors with K-Means
        
        Each word in the vocab represents the centre of a cluster
      - Feature extraction
        
        Generate an unlabelled vocabulary of frequent features
        
        Dense SIFT: PHOW features
        
        Pyramid Histogram Of Words
        
        Dense SIFT - Faster version of SIFT algorithm
        
        4 Scales, uniform spacing
    - - Too small - not representative
      - Too large - quantization artifacts, overfitting