Outline • Announcements • Non-linear filters Announcements • Homework #1 is due today – Please turn in your homework 5/29/2016 Visual Perception Modeling 2 Comments on Edge Detectors • Edge detectors produce good edge maps when the images are piece-wise constant – This is because that edge detectors are assuming that edges are step edges 5/29/2016 Visual Perception Modeling 3 An Example 5/29/2016 Visual Perception Modeling 4 Another Example 5/29/2016 Visual Perception Modeling 5 Real/Natural Images • However, edges may not be very meaningful/useful for real/natural images – Textures – Objects with inhomogeneous colors – Corners 5/29/2016 Visual Perception Modeling 6 An Example Input image 5/29/2016 Canny edge map Visual Perception Modeling 7 Need for Multiple Scales 5/29/2016 Visual Perception Modeling 8 Need for Multiple Scales – cont. 5/29/2016 Visual Perception Modeling 9 Edge Tracking • Edge tracking – Most edges found at large scales tend to be associated with large, high contrast image events • However, the localization is poor at large scales due to smoothing – At fine scales, there are many edges – Edge tracking • Track edges across scales and accept only the fine scale edges that have identifiable parents at a larger scale 5/29/2016 Visual Perception Modeling 10 Problems with Linear Smoothing • To reduce noise, we need to apply some of smoothing – While the smoothing reduces noise, at the same time it also blurs the edges and other important features in the image – At the extreme case, if we apply a Gaussian smoothing filter with a very large , everything will disappear 5/29/2016 Visual Perception Modeling 11 An Example 5/29/2016 Visual Perception Modeling 12 Linear Scale Space • The linear scale space based on the Gaussian kernel can be understood as follows: I ( x, y, t ) I 0 ( x, y) * G( x, y, t ) – where I ( x, y, t ) is the solution of I t I I xx I yy 5/29/2016 Visual Perception Modeling 13 An Example 5/29/2016 Visual Perception Modeling 14 Anisotropic Diffusion • The anisotropic diffusion equation I t div (c( x, y, t )I ) c( x, y, t )I c I – Conductance factor is not uniform any more • Ideally, we would want to encourage smoothing within a region in preference to smoothing across boundaries 5/29/2016 Visual Perception Modeling 15 Anisotropic Diffusion – cont. • The diffusion depends on the local gradient I it,j1 I it, j 5/29/2016 c N N I Visual Perception Modeling 16 Robust Statistics • Non-linear filter as statistical estimator – The goal is to estimate the true value of the pixel in the presence of noisy measurements – This class of filters is extremely useful but very difficult to analyze • Robust estimates – Outliers 5/29/2016 Visual Perception Modeling 17 Median Filters • Given a local neighborhood, the output of the filter is the median of all the values within the neighborhood yij median({ xuv | xuv N }) • Multi-stage filters – The filter responds with the median of a set of different medians, obtained in different neighborhoods 5/29/2016 Visual Perception Modeling 18 Corners and Orientation Representations • Edge detectors fail at corners – The assumption that estimates of the partial derivatives in the x and y direction suffice to estimate an oriented gradient becomes unsupportable • Four types of local windows – – – – Constant windows Edge windows Flow windows 2D windows 5/29/2016 Visual Perception Modeling 19 Corners and Orientation Representations – cont. • Characterization of windows through eigenvalues of the gradient matrix H {(I )(I ) T } window 5/29/2016 Visual Perception Modeling 20