TrainImageContentDetector

✖
`TrainImageContentDetector`

[Experimental]

✖

TrainImageContentDetector[{img₁{bbox₁class₁,…},…}]

trains a ContentDetectorFunction[…] based on the examples given.

Details and Options

TrainImageContentDetector attempts to train an object detector using the provided training data.
Object detection is a computer vision technique used for recognizing subimages in an image.

TrainImageContentDetector returns a ContentDetectorFunction[…] that can then be applied to new input data.
The bounding boxes bbox_i should be given as Rectangle objects.
The classes class_i can be arbitrary expressions.
TrainImageContentDetector[…][image] can be used to detect objects in image and returns the result as a list of associations of the form <|"Image"img,"Class"class|>.
TrainImageContentDetector supports the following options:

	PerformanceGoal	Automatic	favor settings with specific advantages
	ProgressReporting	Automatic	how to report progress during training
	RandomSeeding	1234	how to seed pseudorandom generators internally
	TargetDevice	"CPU"	the target device on which to perform training
	TimeGoal	Automatic	number of seconds to train for
	ValidationSet	None	the set of data on which to evaluate the model during training

Examples

open allclose all

Basic Examples (1)Summary of the most common use cases

Train a basic object detector:

In[105]:=105

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-s45l78

Out[105]=105

Apply the detector on a new image:

In[2]:=2

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-5hl8zj

Out[3]=3

Highlight the detection on the input image:

In[4]:=4

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-z7vn28

Out[4]=4

Options (5)Common values & functionality for each option

PerformanceGoal (1)

Use PerformanceGoal"Quality" to emphasize the quality of the result:

In[1]:=1

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-j1mlnu

Use PerformanceGoal"Speed" to emphasize the speed of computation:

In[3]:=3

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-36wjtx

ProgressReporting (1)

By default, progress is reported in a dynamic panel:

In[1]:=1

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-rrd43o

In[2]:=2

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-tmpaf0

Use ProgressReportingFalse to avoid displaying the progress panel:

In[3]:=3

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-u95rjy

TargetDevice (1)

Train a detector using the default system GPU, if available:

In[4]:=4

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-fz1oph

Out[7]=7

If a compatible GPU is not available, a message is issued:

In[3]:=3

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-3jgboc

Out[3]=3

TimeGoal (1)

The training time can be influenced by several factors, such as the number of examples and classes:

In[3]:=3

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-2gkvvx

In[2]:=2

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-hdnhi8

Use TimeGoal to specify a target time for the training:

In[4]:=4

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-qqme28

ValidationSet (1)

By default, only cross-validation is performed on the detector:

In[8]:=8

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-ljgjr9

In[10]:=10

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-1wpn2m

Out[10]=10

Use ValidationSet to provide separate validation examples:

In[9]:=9

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-0oqk58

In[10]:=10

✖

https://wolfram.com/xid/0bmt8og9plzr1046qxo-eyo3oh

Out[10]=10

Top