ELocatorBase Class
Base class for EasyLocate tool.
The children classes of this class differs by their prediction features (ELocatorFeature). The ground truth objects used to train must have the same set of features as the prediction.
Base Class:EDeepLearningTool
Derived Class(es):EInterestPointLocatorELocator
Namespace: Euresys::Open_eVision::EasyDeepLearning
Methods
Applies the tool on the given image and its mask region.
Evaluates the dataset.
Maximum proximity between two predicted objects regardless of their label.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::AbsoluteMaxOverlap and EInterestPointLocator::AbsoluteMinDistance. Default value: 1.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::AbsoluteMaxOverlap and EInterestPointLocator::AbsoluteMinDistance. Default value: 1.
Capacity of the ELocatorBase.
A higher capacity makes the supervised segmenter capable of learning more information at the cost of a slower processing speed.
A higher capacity makes the supervised segmenter capable of learning more information at the cost of a slower processing speed.
Number of channels of input images. It can either be 1 for grayscale images or 3 for color images.
Detection threshold. The detection threshold is set automatically during training to maximize accuracy. It can also be changed after training to obtain a different true positive/false positive tradeoff.
Height of input images.
The features supported by the locator.
Maximum number of objects in an image (default value: 100).
Prediction anchors.
The prediction anchors are a set of object bounding box sizes. Each anchor is used to detect objects with a size similar to that anchor. As such, the prediction anchors must reflect the variety of sizes of objects that must be detected.
The prediction anchors are a set of object bounding box sizes. Each anchor is used to detect objects with a size similar to that anchor. As such, the prediction anchors must reflect the variety of sizes of objects that must be detected.
Maximum proximity between two predicted objects with the same label.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::SameLabelMaxOverlap and EInterestPointLocator::SameLabelMinDistance. Default value: 0.5
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::SameLabelMaxOverlap and EInterestPointLocator::SameLabelMinDistance. Default value: 0.5
Training metrics at the given iteration
Validation metrics at the given iteration
Width of input images.
Whether the given feature is enabled.
Serializes the settings of the locator.
Maximum proximity between two predicted objects regardless of their label.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::AbsoluteMaxOverlap and EInterestPointLocator::AbsoluteMinDistance. Default value: 1.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::AbsoluteMaxOverlap and EInterestPointLocator::AbsoluteMinDistance. Default value: 1.
Capacity of the ELocatorBase.
A higher capacity makes the supervised segmenter capable of learning more information at the cost of a slower processing speed.
A higher capacity makes the supervised segmenter capable of learning more information at the cost of a slower processing speed.
Number of channels of input images. It can either be 1 for grayscale images or 3 for color images.
Detection threshold. The detection threshold is set automatically during training to maximize accuracy. It can also be changed after training to obtain a different true positive/false positive tradeoff.
Height of input images.
Maximum number of objects in an image (default value: 100).
Prediction anchors.
The prediction anchors are a set of object bounding box sizes. Each anchor is used to detect objects with a size similar to that anchor. As such, the prediction anchors must reflect the variety of sizes of objects that must be detected.
The prediction anchors are a set of object bounding box sizes. Each anchor is used to detect objects with a size similar to that anchor. As such, the prediction anchors must reflect the variety of sizes of objects that must be detected.
Maximum proximity between two predicted objects with the same label.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::SameLabelMaxOverlap and EInterestPointLocator::SameLabelMinDistance. Default value: 0.5
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::SameLabelMaxOverlap and EInterestPointLocator::SameLabelMinDistance. Default value: 0.5
Width of input images.
ELocatorBase Class
Base class for EasyLocate tool.
The children classes of this class differs by their prediction features (ELocatorFeature). The ground truth objects used to train must have the same set of features as the prediction.
Base Class:EDeepLearningTool
Derived Class(es):EInterestPointLocatorELocator
Namespace: Euresys.Open_eVision.EasyDeepLearning
Properties
Maximum proximity between two predicted objects regardless of their label.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::AbsoluteMaxOverlap and EInterestPointLocator::AbsoluteMinDistance. Default value: 1.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::AbsoluteMaxOverlap and EInterestPointLocator::AbsoluteMinDistance. Default value: 1.
Capacity of the ELocatorBase.
A higher capacity makes the supervised segmenter capable of learning more information at the cost of a slower processing speed.
A higher capacity makes the supervised segmenter capable of learning more information at the cost of a slower processing speed.
Number of channels of input images. It can either be 1 for grayscale images or 3 for color images.
Detection threshold. The detection threshold is set automatically during training to maximize accuracy. It can also be changed after training to obtain a different true positive/false positive tradeoff.
Height of input images.
The features supported by the locator.
Maximum number of objects in an image (default value: 100).
Prediction anchors.
The prediction anchors are a set of object bounding box sizes. Each anchor is used to detect objects with a size similar to that anchor. As such, the prediction anchors must reflect the variety of sizes of objects that must be detected.
The prediction anchors are a set of object bounding box sizes. Each anchor is used to detect objects with a size similar to that anchor. As such, the prediction anchors must reflect the variety of sizes of objects that must be detected.
Maximum proximity between two predicted objects with the same label.
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::SameLabelMaxOverlap and EInterestPointLocator::SameLabelMinDistance. Default value: 0.5
The value of the parameter can be between
- 0 when the two objects are completely outside their respective zone of influence; and
- 1 when the two predicted objects are the same.
The actual implementation of the proximity depends on the type of locator tool.
A low value will reduce the amount of falsely detected objects but may increase the number of missed objects. A high value will increase the number of falsely detected objects and reduce the number of missed objects.
See also ELocator::SameLabelMaxOverlap and EInterestPointLocator::SameLabelMinDistance. Default value: 0.5
Width of input images.
Methods
Applies the tool on the given image and its mask region.
Evaluates the dataset.
Training metrics at the given iteration
Validation metrics at the given iteration
Whether the given feature is enabled.
Serializes the settings of the locator.