Main Content

anomalyMapOverlay

Overlay heatmap on image using per-pixel anomaly scores

    Description

    example

    B = anomalyMapOverlay(I,map) fuses an overlay of the per-pixel anomaly scores in map as a heatmap on a grayscale or RGB background image, I.

    Note

    This function requires the Computer Vision Toolbox™ Automated Visual Inspection Library. You can install the Computer Vision Toolbox Automated Visual Inspection Library from Add-On Explorer. For more information about installing add-ons, see Get and Manage Add-Ons. To run this function, you will require the Deep Learning Toolbox™.

    B = anomalyMapOverlay(I,map,Name=Value) specifies options using one or more name-value arguments. For example, anomalyMapOverlay(I,map,Blend="equal") blends the score map and the background image with equal weighting.

    Examples

    collapse all

    Load a pretrained FCDD anomaly detector. This detector has been trained to classify the digit 8 as normal and all other digits as anomalies. Therefore, specify the set of positive labels as the set of digits between 0 and 9, excluding 8.

    load("digit8AnomalyDetector.mat");

    Load test images and corresponding labels, then convert the test images to a formatted dlarray with the format label "SSCB" for spatial-spatial-channel-batch. The data set consists of grayscale images of handwritten digits 0–9.

    [Xtest,catLabels] = digitTest4DArrayData;

    Select a test image of the digit 9.

    digitToTest = "9";
    imTest = Xtest(:,:,:,find(catLabels==digitToTest,1));

    Predict the per-pixel anomaly scores for the test image.

    map = anomalyMap(detector,imTest);

    Display the anomaly map over the test image.

    overlay = anomalyMapOverlay(imTest,map);
    imagesc(overlay)
    axis image off

    Figure contains an axes object. The axes object contains an object of type image.

    Input Arguments

    collapse all

    Background image, specified in one of these formats:

    • M-by-N-by-3 numeric array representing a truecolor image

    • M-by-N numeric matrix representing a grayscale image

    • M-by-N logical matrix representing a binary image

    Anomaly score map, specified as an M-by-N numeric matrix. You can get an anomaly score map using the anomalyMap function. Larger values indicate a greater chance that a pixel is an anomaly.

    Name-Value Arguments

    Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

    Example: B = anomalyMapOverlay(I,map,Blend="equal") blends the score map and the background image with equal weighting.

    Display range of the anomaly score map, map. The default value is [min(map(:)) max(map(:))].

    Weights to blend the map with background image I, specified as "proportional", "equal", or a numeric scalar in the range [0, 1].

    Value Description
    "proportional"Overlay I and map using alpha blending with per-pixel weights proportional to the corresponding values in map. The heatmap is more opaque over regions of interest, which have large anomaly scores. The heatmap is more transparent over normal regions, which have small anomaly scores.
    "equal"Overlay I and map using alpha blending with equal weights for the map and background image.
    numeric scalarOverlay I and map using alpha blending with weight w for the map, where w is the value of the Blend argument. When w is 0, the heatmap is fully transparent and only the background image is visible. When you increase the value of w, the heatmap becomes more opaque.

    For all values of Blend, the weight of the heatmap has an additional multiplicative scale factor, specified by BlendFactor. When you account for the additional scale factor, the weight of the heatmap is w*BlendFactor and the weight of the background image is 1 – w*BlendFactor.

    Constant multiplicative factor to apply to the weights of the Blend argument, specified as a numeric scalar. Increasing the value of BlendFactor increases the brightness of the heatmap over the image.

    Example: BlendFactor=1.2 brightens the heatmap by increasing the weighting by a factor of 1.2.

    Example: BlendFactor=0.6 darkens the heatmap by decreasing the weighting by a factor of 0.6.

    Colormap, specified as a 256-by-3 numeric matrix with values in the range [0, 1]. Each row is a three-element RGB triplet that specifies the red, green, and blue components of a single color of the colormap.

    Data Types: single | double

    Output Arguments

    collapse all

    RGB image with heatmap overlay, returned as an M-by-N-by-3 numeric array.

    Data Types: uint8

    Version History

    Introduced in R2022b