Summary¶

Supported attacks¶

Provides different attack and evaluation approaches.

`CarliniWagnerL2Metric`	The L2 version of C&W attack.
`CarliniWagnerLinfMetric`	The L_inf version of C&W attack.
`AdditiveNoiseMetric`	Base class for metric that tests models against additive noise.
`AdditiveGaussianNoiseMetric`	Metric that tests models against Gaussian noise.
`AdditiveUniformNoiseMetric`	Metric that tests models against uniform noise.
`BlendedUniformNoiseMetric`	Blends the image with a uniform noise image until it is misclassified.
`GaussianBlurMetric`	Metric that tests models against Gaussian blurs.
`BrightnessMetric`	Metric that tests models against brightness variations.
`ContrastReductionMetric`	Metric that tests models against brightness variations.
`MotionBlurMetric`	Motion blurs the image until it is misclassified.
`RotationMetric`	Metric that tests models against rotations.
`SaltAndPepperNoiseMetric`	Add salt and pepper noise.
`SpatialMetric`	Metric that tests models against spatial transformations.

Provides class to wrap existing models in different frameworks so that they provide a unified API to the benchmarks.

`KerasModel`	Create a `Model` instance from a Keras model.
`PyTorchModel`	Creates a `Model` instance from a PyTorch module.
`AipModel`	Base class for models hosted on Baidu AIP platform.
`AipAntiPornModel`	Create a `Model` instance from an AipAntiPorn model.
`GoogleCloudModel`	Base class for models in Google Cloud.
`GoogleSafeSearchModel`	Create a :class: Model instance from a GoogleSafeSearchModel model.
`GoogleObjectDetectionModel`	Create a :class: Model instance from a GoogleObjectDetectionModel model.

`KerasYOLOv3Model`
`KerasSSD300Model`
`KerasResNet50RetinaNetModel`

Provides class to wrap all adversarial criterions so that attacks has uniform API access.

`Misclassification`	Defines adversarials as images for which the predicted class is not the original class.
`ConfidentMisclassification`	Defines adversarials as images for which the probability of any class other than the original is above a given threshold.
`TopKMisclassification`	Defines adversarials as images for which the original class is not one of the top k predicted classes.
`TargetClass`	Defines adversarials as images for which the predicted class is the given target class.
`OriginalClassProbability`	Defines adversarials as images for which the probability of original class is below a given threshold.
`TargetClassProbability`	Defines adversarials as images for which the probability of a given target class is above a given threshold.
`MisclassificationAntiPorn`	Defines adversarials as image for which the probability of being normal is larger than the probability of being porn.
`MisclassificationSafeSearch`	Defines adversarials as image for which the probability of being unsafe is lower than a threshold.
`TargetClassMiss`	Defines adversarials as images for which the target class is not in the detection result.
`TargetClassMissGoogle`	Defines adversarials as images for which the target class is not in the Google object detection result.
`WeightedAP`	Defines adversarials as weighted AP value larger than given threshold.

Provides classes to measure the distance between two images.

`MeanSquaredDistance`	Calculates the mean squared error between two images.
`MeanAbsoluteDistance`	Calculates the mean absolute error between two images.
`Linfinity`	Calculates the L-infinity norm of the difference between two images.
`L0`	Calculates the L0 norm of the difference between two images.
`MSE`	alias of `perceptron.utils.distances.MeanSquaredDistance`
`MAE`	alias of `perceptron.utils.distances.MeanAbsoluteDistance`
`Linf`	alias of `perceptron.utils.distances.Linfinity`