# Documentation

### This is machine translation

Translated by
Mouseover text to see original. Click the button below to return to the English verison of the page.

# trainFasterRCNNObjectDetector

Train a Faster R-CNN deep learning object detector

## Syntax

trainedDetector = trainFasterRCNNObjectDetector(trainingData,network,options)
trainedDetector = trainFasterRCNNObjectDetector(trainingData,checkpoint,options)
trainedDetector = trainFasterRCNNObjectDetector(trainingData,detector,options)
trainedDetector = trainFasterRCNNObjectDetector(___,Name,Value)

## Description

example

trainedDetector = trainFasterRCNNObjectDetector(trainingData,network,options) trains a Faster R-CNN (regions with convolution neural networks) object detector using deep learning. You can train a Faster R-CNN detector to detect multiple object classes. Specify your ground truth training data, your pretrained network, and training options.

This function requires that you have Neural Network Toolbox™. It is recommended that you also have Parallel Computing Toolbox™ to use with a CUDA®-enabled NVIDIA® GPU with compute capability 3.0 or higher.

trainedDetector = trainFasterRCNNObjectDetector(trainingData,checkpoint,options) resumes training from a detector checkpoint.

trainedDetector = trainFasterRCNNObjectDetector(trainingData,detector,options) continues training a detector with additional training data or performs more training iterations to improve detector accuracy.

trainedDetector = trainFasterRCNNObjectDetector(___,Name,Value) uses additional options specified by one or more Name,Value pair arguments and any of the previous inputs.

## Examples

collapse all

trainingData = data.vehicleTrainingData;

trainingData.imageFilename = fullfile(toolboxdir('vision'),'visiondata', ...
trainingData.imageFilename);

Setup network layers.

layers = data.layers
layers =

11x1 Layer array with layers:

1   'imageinput'    Image Input             32x32x3 images with 'zerocenter' normalization
2   'conv_1'        Convolution             32 3x3x3 convolutions with stride [1  1] and padding [1  1]
3   'relu_1'        ReLU                    ReLU
4   'conv_2'        Convolution             32 3x3x32 convolutions with stride [1  1] and padding [1  1]
5   'relu_2'        ReLU                    ReLU
6   'maxpool'       Max Pooling             3x3 max pooling with stride [2  2] and padding [0  0]
7   'fc_1'          Fully Connected         64 fully connected layer
8   'relu_3'        ReLU                    ReLU
9   'fc_2'          Fully Connected         2 fully connected layer
10   'softmax'       Softmax                 softmax
11   'classoutput'   Classification Output   crossentropyex with classes 'vehicle' and 'Background'

Configure training options.

• Lower the InitialLearningRate to reduce the rate at which network parameters are changed.

• Set the CheckpointPath to save detector checkpoints to a temporary directory. Change this to another location if required.

• Set MaxEpochs to 1 to reduce example training time. Increase this to 10 for proper training.

options = trainingOptions('sgdm', ...
'InitialLearnRate', 1e-6, ...
'MaxEpochs', 1, ...
'CheckpointPath', tempdir);

Train detector. Training will take a few minutes.

detector = trainFasterRCNNObjectDetector(trainingData, layers, options)
*************************************************************************
Training a Faster R-CNN Object Detector for the following object classes:

* vehicle

Step 1 of 4: Training a Region Proposal Network (RPN).
|=========================================================================================|
|     Epoch    |   Iteration  | Time Elapsed |  Mini-batch  |  Mini-batch  | Base Learning|
|              |              |  (seconds)   |     Loss     |   Accuracy   |     Rate     |
|=========================================================================================|
|            1 |            1 |         0.18 |       0.7672 |       37.50% |       0.0000 |
|            1 |           50 |         8.18 |       0.7609 |      100.00% |       0.0000 |
|            1 |          100 |        16.75 |       0.6738 |      100.00% |       0.0000 |
|            1 |          150 |        25.25 |       0.4133 |      100.00% |       0.0000 |
|            1 |          200 |        33.82 |       1.2629 |       50.00% |       0.0000 |
|            1 |          250 |        43.45 |       0.6916 |       50.00% |       0.0000 |
|            1 |          295 |        52.21 |       0.2594 |      100.00% |       0.0000 |
|=========================================================================================|

Step 2 of 4: Training a Fast R-CNN Network using the RPN from step 1.
*******************************************************************
Training a Fast R-CNN Object Detector for the following object classes:

* vehicle

--> Extracting region proposals from 295 training images...done.

|=========================================================================================|
|     Epoch    |   Iteration  | Time Elapsed |  Mini-batch  |  Mini-batch  | Base Learning|
|              |              |  (seconds)   |     Loss     |   Accuracy   |     Rate     |
|=========================================================================================|
|            1 |            1 |         0.09 |       0.7288 |       85.71% |       0.0000 |
|            1 |           50 |         3.82 |       0.4308 |       86.96% |       0.0000 |
|            1 |          100 |         7.62 |       0.7009 |       69.57% |       0.0000 |
|            1 |          150 |        11.39 |       0.1261 |       95.24% |       0.0000 |
|            1 |          200 |        15.40 |       0.1939 |      100.00% |       0.0000 |
|            1 |          215 |        16.41 |       0.5769 |       76.19% |       0.0000 |
|=========================================================================================|

Step 3 of 4: Re-training RPN using weight sharing with Fast R-CNN.
|=========================================================================================|
|     Epoch    |   Iteration  | Time Elapsed |  Mini-batch  |  Mini-batch  | Base Learning|
|              |              |  (seconds)   |     Loss     |   Accuracy   |     Rate     |
|=========================================================================================|
|            1 |            1 |         0.17 |       0.5654 |      100.00% |       0.0000 |
|            1 |           50 |         7.56 |       0.3988 |      100.00% |       0.0000 |
|            1 |          100 |        16.09 |       0.2646 |      100.00% |       0.0000 |
|            1 |          150 |        23.77 |       0.8722 |       50.78% |       0.0000 |
|            1 |          200 |        31.83 |       0.3936 |      100.00% |       0.0000 |
|            1 |          250 |        40.20 |       0.4404 |       92.91% |       0.0000 |
|            1 |          295 |        49.23 |       0.7402 |       50.00% |       0.0000 |
|=========================================================================================|

Step 4 of 4: Re-training Fast R-CNN using updated RPN.
*******************************************************************
Training a Fast R-CNN Object Detector for the following object classes:

* vehicle

--> Extracting region proposals from 295 training images...done.

|=========================================================================================|
|     Epoch    |   Iteration  | Time Elapsed |  Mini-batch  |  Mini-batch  | Base Learning|
|              |              |  (seconds)   |     Loss     |   Accuracy   |     Rate     |
|=========================================================================================|
|            1 |            1 |         0.09 |       0.5196 |       72.22% |       0.0000 |
|            1 |           50 |         4.18 |       0.3972 |       72.73% |       0.0000 |
|            1 |          100 |         9.31 |       0.4054 |       87.18% |       0.0000 |
|            1 |          150 |        14.96 |       0.1904 |       96.43% |       0.0000 |
|            1 |          200 |        18.36 |       0.1835 |       96.43% |       0.0000 |
|            1 |          212 |        19.18 |       0.2212 |       95.24% |       0.0000 |
|=========================================================================================|

Finished training Faster R-CNN object detector.

detector =

fasterRCNNObjectDetector with properties:

ModelName: 'vehicle'
Network: [1×1 vision.cnn.FastRCNN]
RegionProposalNetwork: [1×1 vision.cnn.RegionProposalNetwork]
MinBoxSizes: [16 21]
BoxPyramidScale: 2
NumBoxPyramidLevels: 5
ClassNames: {'vehicle'  'Background'}
MinObjectSize: [15 15]

Test the Fast R-CNN detector on a test image.

Run detector.

[bbox, score, label] = detect(detector, img);

Display detection results.

detectedImg = insertShape(img, 'Rectangle', bbox);
figure
imshow(detectedImg)

## Input Arguments

collapse all

Labeled ground truth images, specified as a table with two or more columns. The first column must contain paths and file names to grayscale or truecolor (RGB) images. The remaining columns must contain bounding boxes related to the corresponding image. Each column represents a single object class, such as a car, dog, flower, or stop sign.

Each bounding box must be in the format [x y width height]. The format specifies the upper-left corner location and size of the object in the corresponding image. The table variable name defines the object class name. To create the ground truth table, use the Training Image Labeler app.

Pretrained network, specified as a SeriesNetwork object or as an array of Layer objects. For example:

layers = [imageInputLayer([28 28 3])
convolution2dLayer([5 5],10)
reluLayer()
fullyConnectedLayer(10)
softmaxLayer()
classificationLayer()];

The network is trained to classify the object classes defined in the trainingData table.

When the network is a SeriesNetwork object, the function adjusts the network layers to support the number of object classes defined within the specified trainingData. The background is added as an additional class.

When the network is an array of Layer objects, the network must have a classification layer that supports the number of object classes, plus a background class. Use this input type to customize the learning rates of each layer.

The function replaces the last averagePooling2dLayer or maxPooling2dLayer with an ROI pooling layer.

Training parameters of the neural network, specified using the trainingOptions function. When you specify a single set of training options, the function uses those options for all four training stages. When you specify an array of four options, each stage uses its own set of options. To create an array of four options, assign the trainingOptions function output to each element.

options(1) = trainingOptions('sgdm')
options(2) = trainingOptions('sgdm')
options(3) = trainingOptions('sgdm')
options(4) = trainingOptions('sgdm')

To fine-tune a pre-trained network for detection, lower the initial learning rate to avoid changing the model parameters too rapidly. For example:

options = trainingOptions('sgdm', ...
'InitialLearningRate',1e-6, ...
'CheckpointPath',tempdir);

detector = trainFastRCNNObjectDetector(trainingData,network,options);

To save the detector after every epoch, set the 'CheckpointPath' property when using the trainingOptions function. Saving a checkpoint after every epoch is recommended because network training can take a few hours.

Saved detector checkpoint, specified as a fasterRCNNObjectDetector object. To save the detector after every epoch, set the 'CheckpointPath' property when using the trainingOptions function. Saving a checkpoint after every epoch is recommended because network training can take a few hours.

To load a checkpoint for a previously trained detector, load the MAT-file from the checkpoint path. For example, if the 'CheckpointPath' property of options is '/tmp', load a checkpoint MAT-file using:

The name of the MAT-file includes the iteration number and timestamp of when the detector checkpoint was saved. The detector is saved in the detector variable of the file. Pass this file back into the trainFasterRCNNObjectDetector function:

frcnn = trainFasterRCNNObjectDetector(stopSigns,...
data.detector,options);

Previously trained Faster R-CNN object detector, specified as a fasterRCNNObjectDetector object.

### Name-Value Pair Arguments

Specify optional comma-separated pairs of Name,Value arguments. Name is the argument name and Value is the corresponding value. Name must appear inside single quotes (' '). You can specify several name and value pair arguments in any order as Name1,Value1,...,NameN,ValueN.

Example: 'PositiveOverlapRange',[0.75 1]

collapse all

Bounding box overlap ratios for positive training samples, specified as the comma-separated pair consisting of 'PositiveOverlapRange' and a two-element vector. The vector contains values in the range [0,1]. Region proposals that overlap with ground truth bounding boxes within the specified range are used as positive training samples.

The overlap ratio used for both the PositiveOverlapRange and NegativeOverlapRange is defined as:

$\frac{area\left(A\cap B\right)}{area\left(A\cup B\right)}$

A and B are bounding boxes.

Bounding box overlap ratios for negative training samples, specified as the comma-separated pair consisting of NegativeOverlapRange and a two-element vector. The vector contains values in the range [0,1]. Region proposals that overlap with the ground truth bounding boxes within the specified range are used as negative training samples.

The overlap ratio used for both the PositiveOverlapRange and NegativeOverlapRange is defined as:

$\frac{area\left(A\cap B\right)}{area\left(A\cup B\right)}$

A and B are bounding boxes.

Maximum number of strongest region proposals to use for generating training samples, specified as the comma-separated pair consisting of 'NumStrongestRegions' and a positive integer. Reduce this value to speed up processing time at the cost of training accuracy. To use all region proposals, set this value to Inf.

Length of smallest image dimension, either width or height, specified as the comma-separated pair consisting of 'SmallestImageDimension' and a positive integer. Training images are resized such that the length of the shortest dimension is equal to the specified integer. By default, training images are not resized. Resizing training images helps reduce computational costs and memory used when training images are large. Typical values range from 400–600 pixels.

Minimum anchor box sizes used to build the anchor box pyramid of the region proposal network (RPN), specified as the comma-separated pair consisting of'MinBoxSizes' and an m-by-2 matrix. Each row defines the [height width] of an anchor box.

The default 'auto' setting uses the minimum size and the median aspect ratio from the bounding boxes for each class in the ground truth data. To remove redundant box sizes, the function keeps boxes that have an intersection-over-union that is less than or equal to 0.5. This behavior ensures that the minimum number of anchor boxes are used to cover all the object sizes and aspect ratios.

Anchor box pyramid scale factor used to successively upscale anchor box sizes, specified as the comma-separated pair consisting of 'BoxPyramidScale' and a scalar. Recommended values are from 1 through 2.

Number of levels in an anchor box pyramid, specified as the comma-separated pair consisting of 'NumBoxPyramidLevels' and a scalar. Select a value that ensures that the multiscale anchor boxes are comparable in size to the size of objects in the ground truth data.

The default setting, 'auto', selects the number of levels based on the size of objects within the ground truth data. The number of levels is selected such that it covers the range of object sizes.

## Output Arguments

collapse all

Trained Faster R-CNN object detector, returned as a fasterRCNNObjectDetector object.

## Algorithms

The trainFasterRCNNObjectDetector function uses the alternating training method [1].

## References

[1] Ren, Shaoqing, Kaiming He, Ross Girshick, and Jian Sun. "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks." Advances in Neural Information Processing Systems . Vol. 28, 2015.