balanceBoxLabels

Balance bounding box labels for object detection

Since R2020a

Syntax

locationSet = balanceBoxLabels(boxLabels,blockedImages,blockSize,numObservations)

locationSet = balanceBoxLabels(boxLabels,blockedImages,blockSize,numObservations,Name,Value)

Description

locationSet = balanceBoxLabels(boxLabels,blockedImages,blockSize,numObservations) balances bounding box labels, boxLabels, by oversampling blocks of images containing less frequent classes, contained in the collection of blocked image objects blockedImages. numObservations is the required number of block locations, and blockSize specifies the block size.

locationSet = balanceBoxLabels(boxLabels,blockedImages,blockSize,numObservations,Name,Value) specifies additional aspects of the selected blocks using name-value arguments.

Examples

Sample Block Sets to Use in Blocked Image Object Detection

Load box labels data that contains boxes and labels for one image. The height and width of each box is [20,20].

d  = load('balanceBoxLabelsData.mat');
boxLabels    = d.BoxLabels;

Create a blocked image of size [500,500].

blockedImages = blockedImage(zeros([500,500]));

Choose the images size of each observation.

blockSize = [50,50];

Visualize using a histogram to identify any class imbalance in the box labels.

blds = boxLabelDatastore(boxLabels);
datasetCount = countEachLabel(blds);
figure;
h1 = histogram('Categories',datasetCount.Label,'BinCounts',datasetCount.Count)

h1 = 
  Histogram with properties:

              Data: [0x0 categorical]
            Values: [1 1 1 1 1 1 1 1 1 1 1 11]
    NumDisplayBins: 12
        Categories: {'1'  '2'  '3'  '4'  '5'  '6'  '7'  '8'  '9'  '10'  '11'  '12'}
      DisplayOrder: 'manual'
     Normalization: 'count'
      DisplayStyle: 'bar'
         FaceColor: 'auto'
         EdgeColor: [0 0 0]

  Show all properties

Measure the distribution of box labels. If the coefficent of variation is more than 1, then there is class imbalance.

cvBefore = std(datasetCount.Count)/mean(datasetCount.Count)

cvBefore = 1.5746

Choose a heuristic value for number of observations by finding the mean of the counts of each class, multiplied by the number of classes.

numClasses = height(datasetCount);
numObservations = mean(datasetCount.Count) * numClasses;

Control the amount a box can be cut using OverlapThreshold. Using a lower threshold value will cut objects more at the border of a block. Increase this value to reduce the amount an object can be clipped at the border, at the expense of a less balanced box labels.

ThresholdValue = 0.5;

Balance boxLabels using the balanceBoxLabels function.

locationSet = balanceBoxLabels(boxLabels,blockedImages,blockSize,...
        numObservations,'OverlapThreshold',ThresholdValue);

[==================================================] 100%
Elaps[==================================================] 100%
Elapsed time: 00:00:00
Estimated time remaining: 00:00:00
Balancing box labels complete.

Count the labels that are contained within the image blocks.

bldsBalanced = boxLabelDatastore(boxLabels,locationSet);
balancedDatasetCount = countEachLabel(bldsBalanced);

Overlay another histogram against the original label count to see if the box labels are balanced. If the labels appear to be not balanced by looking at the histograms, increase the value for numObservations.

hold on;
balancedLabels = balancedDatasetCount.Label;
balancedCount  = balancedDatasetCount.Count;
h2 = histogram('Categories',balancedLabels,'BinCounts',balancedCount);
title(h2.Parent,"Balanced class labels (OverlapThreshold: " + ThresholdValue + ")" );
legend(h2.Parent,{'Before','After'});

Measure the distribution of the new baanced box labels.

cvAfter = std(balancedCount)/mean(balancedCount)

cvAfter = 0.4588

Input Arguments

boxLabels — Labeled bounding box data

table with two columns

Labeled bounding box data, specified as a table with two columns.

• The first column contains bounding boxes and must be a cell vector. Each element in the cell vector contains M-by-4 matrices in the format [x, y, width, height] for M boxes.

• The second column must be a cell vector that contains the label names corresponding to each bounding box. Each element in the cell vector must be an M-by-1 categorical or string vector.

  

To create a box label table from ground truth data,

1. Use the Image Labeler or Video Labeler app to label your ground truth. Export the labeled ground truth data to your workspace.

2. Create a bounding box label datastore using the objectDetectorTrainingData function.

3. You can obtain the boxLabels from the LabelData property of the box label datastore returned by objectDetectorTrainingData, ( blds.LabelData).

blockedImages — Labeled blocked images

array of blockedImage objects

Labeled blocked images, specified as an array of blockedImage objects containing pixel label images.

blockSize — Block size

two-element row vector of positive integers

Block size of read data, specified as a two-element row vector of positive integers, [numrows,numcols]. The first element specifies the number of rows in the block. The second element specifies the number of columns.

numObservations — Number of block locations

positive integer

Number of block locations to return, specified as a positive integer.

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.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: 'OverlapThreshold','1'

Levels — Resolution level of each image

1 (default) | positive integer scalar | B-by-1 vector of positive integers

Resolution level of each image in the array of blockedImage objects, specified as a positive integer scalar or a B-by-1 vector of positive integers, where B is the length of the array of blockedImage objects.

OverlapThreshold — Overlap threshold

1 (default) | scalar in the range [0,1]

Overlap threshold, specified as a positive scalar in the range [0,1]. When the overlap between a bounding box and a cropping window is greater than the threshold, boxes in the boxLabels input are clipped to the image block window border. When the overlap is less than the threshold, the boxes are discarded. When you lower the threshold, part of an object can get discarded. To reduce the amount an object can be clipped at the border, increase the threshold. Increasing the threshold can also cause less-balanced box labels.

The amount of overlap between the bounding box and a cropping window is defined as.

$$area(bboxA\cap window)/area(bboxA)$$

Verbose — Display progress information

true or 1 (default) | false or 0

Display progress information, specified as a numeric or logical 1 (true) or 0 (false). Set this property to true to display information.

Output Arguments

locationSet — Balanced box labels

blockLocationSet object

Balanced box labels, returned as a blockLocationSet object. The object contains numObservations number of locations of balanced blocks, each of size blockSize.

Algorithms

Balancing Box Labels

To balance box labels, the function over samples classes that are less represented in the blocked image or big image. The box labels are counted across the dataset and sorted based on each class count. Each image size is split into several quadrants, based on the blockSize input value. The algorithm randomly picks several blocks within each quadrant with less-represented classes. The blocks without any objects are discarded. The balancing stops once the specified number of blocks are selected.

Checking for Balance

You can check the success of balancing by comparing the histograms of label count before and after balancing. You can also check the coefficient of variation value. For best results, the value should be less than the original value. For more information, see the National Institute of Standards and Technology (NIST) website, see Coefficient of Variation for more information.

Version History

Introduced in R2020a

R2021a: bigLabeledImages argument is not recommended

The bigLabeledImages argument, which supports bigimage objects, is not recommended. Use the blockedImages argument instead, which supports blockedImage objects. The blockedImage object offers several advantages including extension to N-D processing, a simpler interface, and custom support for reading and writing nonstandard image formats.

Although there are no plans to remove the bigLabeledImages argument at this time, switch to the blockedImages argument to take advantage of the additional capabilities and flexibility.

To update your code, follow these steps:

• Replace bigimage object input with blockedImage object input for the second argument of this function.

• If you want to select blocks of any of the blocked images at a resolution level other than 1, then specify the 'Levels' name-value argument. You can omit this argument when you want to select blocks from all blocked images at resolution level 1.

The table gives an example of how to update your code.

Discouraged Usage	Recommended Replacement
This example selects blocks at resolution level 1 from a bigimage object. boxLabels = load('balanceBoxLabelsData.mat').BoxLabels; bim = bigimage(zeros([500,500])); blockSize = [50 50]; numObservations = 20; locationSet = balanceBoxLabels(boxLabels,bim,1, ... blockSize,numObservations);	Here is equivalent code, replacing the input bigimage object with a blockedImage object. boxLabels = load('balanceBoxLabelsData.mat').BoxLabels; bim = blockedImage(zeros([500,500])); blockSize = [50 50]; numObservations = 20; locationSet = balanceBoxLabels(boxLabels,bim, ... blockSize,numObservations);