anonymous_namespace{Net_.cpp} Namespace Reference

Functions
MatND	MxArrayToBlob (const MxArray &arr)
	Create 4-dimensional blob from MATLAB array. More...
Net::LayerId	MxArrayToLayerId (const MxArray &arr)
	Convert MxArray to cv::dnn::Net::LayerId. More...
vector< Net::LayerId >	MxArrayToVectorLayerId (const MxArray &arr)
	Convert MxArray to std::vector<cv::dnn::Net::LayerId> More...
LayerParams	MxArrayToLayerParams (const MxArray &arr)
	Convert MxArray to cv::dnn::LayerParams. More...
MxArray	toStruct (const Ptr< Layer > &layer)
	Convert cv::Ptr<cv::dnn::Layer> to scalar struct. More...
MxArray	toStruct (const vector< Ptr< Layer > > &layers)
	Convert std::vector<cv::Ptr<cv::dnn::Layer>> to struct array. More...
MxArray	toMxArray (int64_t i)
	MxArray constructor from 64-bit integer. More...
Ptr< Net >	readNetFrom (const string &type, vector< MxArray >::const_iterator first, vector< MxArray >::const_iterator last)
	Create an instance of Net using options in arguments. More...

Variables
int	last_id = 0
	Last object id to allocate. More...
map< int, Ptr< Net > >	obj_
	Object container. More...
const ConstMap< string, int >	BackendsMap
	Computation backends for option processing. More...
const ConstMap< string, int >	TargetsMap
	Computation target devices for option processing. More...
const ConstMap< int, string >	TargetsInvMap
	Computation target devices for option processing. More...

Function Documentation

MxArrayToBlob()

MatND anonymous_namespace{Net_.cpp}::MxArrayToBlob

(

const MxArray &

arr

)

Create 4-dimensional blob from MATLAB array.

Parameters

arr	input MxArray object (numeric array).

Returns

blob 4-dimensional cv::MatND.

See also

MxArray::toMatND

Definition at line 43 of file Net_.cpp.

MxArrayToLayerId()

Net::LayerId anonymous_namespace{Net_.cpp}::MxArrayToLayerId

(

const MxArray &

arr

)

Convert MxArray to cv::dnn::Net::LayerId.

Parameters

arr	MxArray object. In one of the following forms: a scalar integer. a scalar double. a string.

Returns

instance of LayerId struct (a typedef for cv::dnn::DictValue, which is a container for either a string, a double, or an integer.

Definition at line 65 of file Net_.cpp.

MxArrayToLayerParams()

LayerParams anonymous_namespace{Net_.cpp}::MxArrayToLayerParams

(

const MxArray &

arr

)

Convert MxArray to cv::dnn::LayerParams.

Parameters

arr

MxArray object. A scalar struct with the following fields: dict: Scalar struct of key/value dictionary parameters. blobs: Cell array of learned parameters stored as blobs. name: Name of the layer instance (optional, used for internal purposes). type: Type name which was used for creating layer by factory (optional).

Returns

instance of created LayerParams struct.

Definition at line 114 of file Net_.cpp.

MxArrayToVectorLayerId()

vector<Net::LayerId> anonymous_namespace{Net_.cpp}::MxArrayToVectorLayerId

(

const MxArray &

arr

)

Convert MxArray to std::vector<cv::dnn::Net::LayerId>

Parameters

arr	MxArray object. Ine one of the following forms: a cell array of scalars (integers of doubles) a cell array of strings a numeric array of integers or doubles

Returns

vector of LayerId

Definition at line 82 of file Net_.cpp.

readNetFrom()

Ptr<Net> anonymous_namespace{Net_.cpp}::readNetFrom	(	const string &	type,
		vector< MxArray >::const_iterator	first,
		vector< MxArray >::const_iterator	last
	)

Create an instance of Net using options in arguments.

Parameters

type	type of network to import, one of: "Caffe" "Tensorflow" "Torch" "Darknet"
first	iterator at the beginning of the vector range
last	iterator at the end of the vector range

Returns

smart pointer to created Net

Definition at line 216 of file Net_.cpp.

toMxArray()

MxArray anonymous_namespace{Net_.cpp}::toMxArray

(

int64_t

i

)

MxArray constructor from 64-bit integer.

Parameters

i	int value.

Returns

MxArray object, a scalar int64 array.

Definition at line 197 of file Net_.cpp.

toStruct() [1/2]

MxArray anonymous_namespace{Net_.cpp}::toStruct

(

const Ptr< Layer > &

layer

)

Convert cv::Ptr<cv::dnn::Layer> to scalar struct.

Parameters

layer

smart pointer to an instance of Layer

Returns

scalar struct MxArray object

Definition at line 165 of file Net_.cpp.

toStruct() [2/2]

MxArray anonymous_namespace{Net_.cpp}::toStruct

(

const vector< Ptr< Layer > > &

layers

)

Convert std::vector<cv::Ptr<cv::dnn::Layer>> to struct array.

Parameters

layers

vector of smart pointers to layers

Returns

struct-array MxArray object

Definition at line 180 of file Net_.cpp.

Variable Documentation

BackendsMap

const ConstMap<string,int> anonymous_namespace{Net_.cpp}::BackendsMap

Initial value:

= ConstMap<string,int>
    ("Default",         cv::dnn::DNN_BACKEND_DEFAULT)
    ("Halide",          cv::dnn::DNN_BACKEND_HALIDE)
    ("InferenceEngine", cv::dnn::DNN_BACKEND_INFERENCE_ENGINE)

Computation backends for option processing.

Definition at line 22 of file Net_.cpp.

last_id

int anonymous_namespace{Net_.cpp}::last_id = 0

Last object id to allocate.

Definition at line 17 of file Net_.cpp.

obj_

map<int,Ptr<Net> > anonymous_namespace{Net_.cpp}::obj_

Object container.

Definition at line 19 of file Net_.cpp.

TargetsInvMap

const ConstMap<int,string> anonymous_namespace{Net_.cpp}::TargetsInvMap

Initial value:

= ConstMap<int,string>
    (cv::dnn::DNN_TARGET_CPU,    "CPU")
    (cv::dnn::DNN_TARGET_OPENCL, "OpenCL")

Computation target devices for option processing.

Definition at line 33 of file Net_.cpp.

TargetsMap

const ConstMap<string,int> anonymous_namespace{Net_.cpp}::TargetsMap

Initial value:

= ConstMap<string,int>
    ("CPU",    cv::dnn::DNN_TARGET_CPU)
    ("OpenCL", cv::dnn::DNN_TARGET_OPENCL)

Computation target devices for option processing.

Definition at line 28 of file Net_.cpp.