countEachLabel

Count occurrence of pixel or box labels

Syntax

counts = countEachLabel(ds)

Description

counts = countEachLabel(ds) returns a table containing information about the pixel or box labels and count for the input datastore, ds.

Examples

Count Unique Labels in Datastore

Load a table that contains bounding boxes with labels for vehicles.

load('vehicleTrainingData.mat');

Load a table that contains bounding boxes with labels for stop signs and cars.

load('stopSignsAndCars.mat');

Combine ground truth boxes and labels, excluding the image filenames in the first column.

vehiclesTbl  = vehicleTrainingData(:,2:end);
stopSignsTbl = stopSignsAndCars(:,2:end);

Create a boxLabelDatastore using 2 tables, one with vehicle label data and the other with stop signs label data.

blds = boxLabelDatastore(vehiclesTbl,stopSignsTbl);
tbl = countEachLabel(blds)

tbl=4×3 table
     Label      Count    ImageCount
    ________    _____    __________

    vehicle      336        295    
    stopSign      42         41    
    carRear       10          9    
    carFront       9          8    

Create a histogram plot using the labels and the respective label counts.

histogram('Categories',tbl.Label,'BinCounts',tbl.Count);

Create another histogram overlaying the respective image counts.

hold on;
histogram('Categories',tbl.Label,'BinCounts',tbl.ImageCount);

Pass Class Weights to Pixel Classification Layer

Set the location of image and pixel label data.

dataDir = fullfile(toolboxdir('vision'),'visiondata');
imDir = fullfile(dataDir,'building');
pxDir = fullfile(dataDir,'buildingPixelLabels');

Create a pixel label image datastore using the ground truth images in imds and the pixel labeled images in pxds.

imds = imageDatastore(imDir);
classNames = ["sky" "grass" "building" "sidewalk"];
pixelLabelID = [1 2 3 4];
pxds = pixelLabelDatastore(pxDir,classNames,pixelLabelID);

Tabulate pixel label counts in dataset.

tbl = countEachLabel(pxds)

tbl=4×3 table
        Name        PixelCount    ImagePixelCount
    ____________    __________    _______________

    {'sky'     }    3.1485e+05       1.536e+06   
    {'grass'   }    1.5979e+05       1.536e+06   
    {'building'}    1.0312e+06       1.536e+06   
    {'sidewalk'}         25313       9.216e+05   

Balance classes using uniform prior weighting.

prior = 1/numel(classNames);
uniformClassWeights = prior./tbl.PixelCount

uniformClassWeights = 4×1
10-5 ×

    0.0794
    0.1565
    0.0242
    0.9876

Balance classes using inverse frequency weighting.

totalNumberOfPixels = sum(tbl.PixelCount);
frequency = tbl.PixelCount / totalNumberOfPixels;
invFreqClassWeights = 1./frequency

invFreqClassWeights = 4×1

    4.8632
    9.5827
    1.4848
   60.4900

Balance classes using median frequency weighting.

freq = tbl.PixelCount ./ tbl.ImagePixelCount

freq = 4×1

    0.2050
    0.1040
    0.6714
    0.0275

medFreqClassWeights = median(freq) ./ freq

medFreqClassWeights = 4×1

    0.7538
    1.4852
    0.2301
    5.6252

Pass the class weights using median frequency weighting to the pixel classification layer.

layer = pixelClassificationLayer('Classes',tbl.Name, ...
  'ClassWeights', medFreqClassWeights)

layer = 
  PixelClassificationLayer with properties:

            Name: ''
         Classes: [sky    grass    building    sidewalk]
    ClassWeights: [4x1 double]
      OutputSize: 'auto'

   Hyperparameters
    LossFunction: 'crossentropyex'

Input Arguments

ds — Datastore with labeled data

PixelLabelDatastore object | pixelLabelImageDatastore object | boxLabelDatastore object

Datastore with labeled data for training a semantic segmentation network or an object detection network, specified as a pixelLabelDatastore, pixelLabelImageDatastore, or boxLabelDatastore object.

Output Arguments

counts — Label information

table

Label information, returned as a table. The labeled data table contain three variables.

For pixelLabelDatastore and pixelLabelImageDatastore inputs, the counts output contains:

Pixel Count Variables	Description
Name	Pixel label class name
PixelCount	Number of pixels in class
ImagePixelCount	Total number of pixels in images that had an instance of the class

For boxLabelDatastore inputs, the counts output table contains:

Box Count Variables	Description
Label	Box label class name
Count	Total number of labels of the class across all images
ImageCount	Total number of images that contain one or more instances of the class

Tips

The output of countEachLabel can be used to calculate class weights for class balancing. For example, for labeled pixel data information in tbl:

• Uniform class balancing weights each class such that each contains a uniform prior probability:

  numClasses = height(tbl)
  prior = 1/numClasses;
  classWeights = prior./tbl.PixelCount

• Inverse frequency balancing weights each class such that underrepresented classes are given higher weight:

  totalNumberOfPixels = sum(tbl.PixelCount)
  frequency = tbl.PixelCount / totalNumberOfPixels;
  classWeights = 1./frequency

• Median frequency balancing weights each class using the median frequency. The weight for each class is defined as median(imageFreq)/imageFreq(c), where imageFreq(c) represents the number of pixels of the class divided by the total number of pixels in images that had an instance of the class (c):

  imageFreq = tbl.PixelCount ./ tbl.ImagePixelCount
  classWeights = median(imageFreq) ./ imageFreq
  

  The calculated class weights can be passed to the pixelClassificationLayer.

Version History

Introduced in R2017b