Class Net

This class allows to create and manipulate comprehensive artificial neural networks.

Inheritance
System.Object
DisposableObject
DisposableCvObject
Net
Implements
Inherited Members
Namespace: OpenCvSharp.Dnn
Assembly: OpenCvSharp.dll
Syntax
public class Net : DisposableCvObject, ICvPtrHolder
Remarks

Neural network is presented as directed acyclic graph(DAG), where vertices are Layer instances, and edges specify relationships between layers inputs and outputs.

Each network layer has unique integer id and unique string name inside its network. LayerId can store either layer name or layer id. This class supports reference counting of its instances, i.e.copies point to the same instance.

Constructors

| Improve this Doc View Source

Net()

Default constructor.

Declaration
public Net()
| Improve this Doc View Source

Net(IntPtr)

Declaration
protected Net(IntPtr ptr)
Parameters
Type Name Description
IntPtr ptr

Methods

| Improve this Doc View Source

Connect(Int32, Int32, Int32, Int32)

Connects #@p outNum output of the first layer to #@p inNum input of the second layer.

Declaration
public void Connect(int outLayerId, int outNum, int inpLayerId, int inpNum)
Parameters
Type Name Description
System.Int32 outLayerId

identifier of the first layer
System.Int32 outNum

identifier of the second layer
System.Int32 inpLayerId

number of the first layer output
System.Int32 inpNum

number of the second layer input
| Improve this Doc View Source

Connect(String, String)

Connects output of the first layer to input of the second layer.

Declaration
public void Connect(string outPin, string inpPin)
Parameters
Type Name Description
System.String outPin

descriptor of the first layer output.
System.String inpPin

descriptor of the second layer input.
| Improve this Doc View Source

DisposeUnmanaged()

Declaration
protected override void DisposeUnmanaged()
Overrides
| Improve this Doc View Source

Dump()

Dump net to String. Call method after setInput(). To see correct backend, target and fusion run after forward().

Declaration
public string Dump()
Returns
Type Description
System.String

String with structure, hyperparameters, backend, target and fusion
| Improve this Doc View Source

DumpToFile(String)

Dump net structure, hyperparameters, backend, target and fusion to dot file

Declaration
public void DumpToFile(string path)
Parameters
Type Name Description
System.String path

path to output file with .dot extension
| Improve this Doc View Source

Empty()

Returns true if there are no layers in the network.

Declaration
public bool Empty()
Returns
Type Description
System.Boolean
| Improve this Doc View Source

EnableFusion(Boolean)

Enables or disables layer fusion in the network.

Declaration
public void EnableFusion(bool fusion)
Parameters
Type Name Description
System.Boolean fusion

true to enable the fusion, false to disable. The fusion is enabled by default.
| Improve this Doc View Source

Forward(IEnumerable<Mat>, IEnumerable<String>)

Runs forward pass to compute outputs of layers listed in @p outBlobNames.

Declaration
public void Forward(IEnumerable<Mat> outputBlobs, IEnumerable<string> outBlobNames)
Parameters
Type Name Description
IEnumerable<Mat> outputBlobs

contains blobs for first outputs of specified layers.
IEnumerable<System.String> outBlobNames

names for layers which outputs are needed to get
| Improve this Doc View Source

Forward(IEnumerable<Mat>, String)

Runs forward pass to compute output of layer with name @p outputName.

Declaration
public void Forward(IEnumerable<Mat> outputBlobs, string outputName = null)
Parameters
Type Name Description
IEnumerable<Mat> outputBlobs

contains all output blobs for specified layer.
System.String outputName

name for layer which output is needed to get. If outputName is empty, runs forward pass for the whole network.
| Improve this Doc View Source

Forward(String)

Runs forward pass to compute output of layer with name @p outputName. By default runs forward pass for the whole network.

Declaration
public Mat Forward(string outputName = null)
Parameters
Type Name Description
System.String outputName

name for layer which output is needed to get
Returns
Type Description
Mat

blob for first output of specified layer.
| Improve this Doc View Source

GetLayerId(String)

Converts string name of the layer to the integer identifier.

Declaration
public int GetLayerId(string layer)
Parameters
Type Name Description
System.String layer
Returns
Type Description
System.Int32

id of the layer, or -1 if the layer wasn't found.
| Improve this Doc View Source

GetLayerNames()

Declaration
public string[] GetLayerNames()
Returns
Type Description
System.String[]
| Improve this Doc View Source

GetPerfProfile(out Double[])

Returns overall time for inference and timings (in ticks) for layers. Indexes in returned vector correspond to layers ids.Some layers can be fused with others, in this case zero ticks count will be return for that skipped layers.

Declaration
public long GetPerfProfile(out double[] timings)
Parameters
Type Name Description
System.Double[] timings

vector for tick timings for all layers.
Returns
Type Description
System.Int64

overall ticks for model inference.
| Improve this Doc View Source

GetUnconnectedOutLayers()

Returns indexes of layers with unconnected outputs.

Declaration
public int[] GetUnconnectedOutLayers()
Returns
Type Description
System.Int32[]
| Improve this Doc View Source

GetUnconnectedOutLayersNames()

Returns names of layers with unconnected outputs.

Declaration
public string[] GetUnconnectedOutLayersNames()
Returns
Type Description
System.String[]
| Improve this Doc View Source

ReadFromModelOptimizer(String, String)

Create a network from Intel's Model Optimizer intermediate representation (IR). Networks imported from Intel's Model Optimizer are launched in Intel's Inference Engine backend.

Declaration
public static Net ReadFromModelOptimizer(string xml, string bin)
Parameters
Type Name Description
System.String xml

XML configuration file with network's topology.
System.String bin

Binary file with trained weights.
Returns
Type Description
Net
| Improve this Doc View Source

ReadNet(String, String, String)

Read deep learning network represented in one of the supported formats.

This function automatically detects an origin framework of trained model and calls an appropriate function such @ref readNetFromCaffe, @ref readNetFromTensorflow,

Declaration
public static Net ReadNet(string model, string config = "", string framework = "")
Parameters
Type Name Description
System.String model

Binary file contains trained weights. The following file

  •              extensions are expected for models from different frameworks:

  •              * `*.caffemodel` (Caffe, http://caffe.berkeleyvision.org/)

  •              * `*.pb` (TensorFlow, https://www.tensorflow.org/)

  •              * `*.t7` | `*.net` (Torch, http://torch.ch/)

  •              * `*.weights` (Darknet, https://pjreddie.com/darknet/)

  •              * `*.bin` (DLDT, https://software.intel.com/openvino-toolkit)
System.String config

Text file contains network configuration. It could be a

  •               file with the following extensions:

  •              * `*.prototxt` (Caffe, http://caffe.berkeleyvision.org/)

  •              * `*.pbtxt` (TensorFlow, https://www.tensorflow.org/)

  •              * `*.cfg` (Darknet, https://pjreddie.com/darknet/)

  •              * `*.xml` (DLDT, https://software.intel.com/openvino-toolkit)
System.String framework

Explicit framework name tag to determine a format.
Returns
Type Description
Net
| Improve this Doc View Source

ReadNetFromCaffe(ReadOnlySpan<Byte>, ReadOnlySpan<Byte>)

Reads a network model stored in Caffe model files from memory.

Declaration
public static Net ReadNetFromCaffe(ReadOnlySpan<byte> bufferProto, ReadOnlySpan<byte> bufferModel = null)
Parameters
Type Name Description
ReadOnlySpan<System.Byte> bufferProto

buffer containing the content of the .prototxt file
ReadOnlySpan<System.Byte> bufferModel

buffer containing the content of the .caffemodel file
Returns
Type Description
Net
Remarks

This is shortcut consisting from createCaffeImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromCaffe(Byte[], Byte[])

Reads a network model stored in Caffe model in memory.

Declaration
public static Net ReadNetFromCaffe(byte[] bufferProto, byte[] bufferModel = null)
Parameters
Type Name Description
System.Byte[] bufferProto

buffer containing the content of the .prototxt file
System.Byte[] bufferModel

buffer containing the content of the .caffemodel file
Returns
Type Description
Net
Remarks

This is shortcut consisting from createCaffeImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromCaffe(String, String)

Reads a network model stored in Caffe model files.

Declaration
public static Net ReadNetFromCaffe(string prototxt, string caffeModel = null)
Parameters
Type Name Description
System.String prototxt

path to the .prototxt file with text description of the network architecture.
System.String caffeModel

path to the .caffemodel file with learned network.
Returns
Type Description
Net
Remarks

This is shortcut consisting from createCaffeImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromDarknet(ReadOnlySpan<Byte>, ReadOnlySpan<Byte>)

Reads a network model stored in Caffe model files from memory.

Declaration
public static Net ReadNetFromDarknet(ReadOnlySpan<byte> bufferCfg, ReadOnlySpan<byte> bufferModel = null)
Parameters
Type Name Description
ReadOnlySpan<System.Byte> bufferCfg

A buffer contains a content of .cfg file with text description of the network architecture.
ReadOnlySpan<System.Byte> bufferModel

A buffer contains a content of .weights file with learned network.
Returns
Type Description
Net
Remarks

This is shortcut consisting from createCaffeImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromDarknet(Byte[], Byte[])

Reads a network model stored in Caffe model files from memory.

Declaration
public static Net ReadNetFromDarknet(byte[] bufferCfg, byte[] bufferModel = null)
Parameters
Type Name Description
System.Byte[] bufferCfg

A buffer contains a content of .cfg file with text description of the network architecture.
System.Byte[] bufferModel

A buffer contains a content of .weights file with learned network.
Returns
Type Description
Net
Remarks

This is shortcut consisting from createCaffeImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromDarknet(String, String)

Reads a network model stored in Darknet (https://pjreddie.com/darknet/) model files.

Declaration
public static Net ReadNetFromDarknet(string cfgFile, string darknetModel = null)
Parameters
Type Name Description
System.String cfgFile

path to the .cfg file with text description of the network architecture.
System.String darknetModel

path to the .weights file with learned network.
Returns
Type Description
Net

Network object that ready to do forward, throw an exception in failure cases.
Remarks

This is shortcut consisting from DarknetImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromModelOptimizer(String, String)

Load a network from Intel's Model Optimizer intermediate representation. Networks imported from Intel's Model Optimizer are launched in Intel's Inference Engine backend.

Declaration
public static Net ReadNetFromModelOptimizer(string xml, string bin)
Parameters
Type Name Description
System.String xml

XML configuration file with network's topology.
System.String bin

Binary file with trained weights.
Returns
Type Description
Net
| Improve this Doc View Source

ReadNetFromONNX(ReadOnlySpan<Byte>)

Reads a network model ONNX https://onnx.ai/ from memory

Declaration
public static Net ReadNetFromONNX(ReadOnlySpan<byte> onnxFileData)
Parameters
Type Name Description
ReadOnlySpan<System.Byte> onnxFileData

memory of the first byte of the buffer.
Returns
Type Description
Net

Network object that ready to do forward, throw an exception in failure cases.
| Improve this Doc View Source

ReadNetFromONNX(Byte[])

Reads a network model ONNX https://onnx.ai/ from memory

Declaration
public static Net ReadNetFromONNX(byte[] onnxFileData)
Parameters
Type Name Description
System.Byte[] onnxFileData

memory of the first byte of the buffer.
Returns
Type Description
Net

Network object that ready to do forward, throw an exception in failure cases.
| Improve this Doc View Source

ReadNetFromONNX(String)

Reads a network model ONNX https://onnx.ai/

Declaration
public static Net ReadNetFromONNX(string onnxFile)
Parameters
Type Name Description
System.String onnxFile

path to the .onnx file with text description of the network architecture.
Returns
Type Description
Net

Network object that ready to do forward, throw an exception in failure cases.
| Improve this Doc View Source

ReadNetFromTensorflow(ReadOnlySpan<Byte>, ReadOnlySpan<Byte>)

Reads a network model stored in Tensorflow model from memory.

Declaration
public static Net ReadNetFromTensorflow(ReadOnlySpan<byte> bufferModel, ReadOnlySpan<byte> bufferConfig = null)
Parameters
Type Name Description
ReadOnlySpan<System.Byte> bufferModel

buffer containing the content of the pb file
ReadOnlySpan<System.Byte> bufferConfig

buffer containing the content of the pbtxt file (optional)
Returns
Type Description
Net
Remarks

This is shortcut consisting from createTensorflowImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromTensorflow(Byte[], Byte[])

Reads a network model stored in Tensorflow model from memory.

Declaration
public static Net ReadNetFromTensorflow(byte[] bufferModel, byte[] bufferConfig = null)
Parameters
Type Name Description
System.Byte[] bufferModel

buffer containing the content of the pb file
System.Byte[] bufferConfig

buffer containing the content of the pbtxt file (optional)
Returns
Type Description
Net
Remarks

This is shortcut consisting from createTensorflowImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromTensorflow(String, String)

Reads a network model stored in Tensorflow model file.

Declaration
public static Net ReadNetFromTensorflow(string model, string config = null)
Parameters
Type Name Description
System.String model

path to the .pb file with binary protobuf description of the network architecture
System.String config

path to the .pbtxt file that contains text graph definition in protobuf format.
Returns
Type Description
Net

Resulting Net object is built by text graph using weights from a binary one that let us make it more flexible.
Remarks

This is shortcut consisting from createTensorflowImporter and Net::populateNet calls.

| Improve this Doc View Source

ReadNetFromTorch(String, Boolean)

Reads a network model stored in Torch model file.

Declaration
public static Net ReadNetFromTorch(string model, bool isBinary = true)
Parameters
Type Name Description
System.String model
System.Boolean isBinary
Returns
Type Description
Net
Remarks

This is shortcut consisting from createTorchImporter and Net::populateNet calls.

| Improve this Doc View Source

SetHalideScheduler(String)

Compile Halide layers. Schedule layers that support Halide backend. Then compile them for specific target.For layers that not represented in scheduling file or if no manual scheduling used at all, automatic scheduling will be applied.

Declaration
public void SetHalideScheduler(string scheduler)
Parameters
Type Name Description
System.String scheduler

Path to YAML file with scheduling directives.
| Improve this Doc View Source

SetInput(Mat, String)

Sets the new value for the layer output blob

Declaration
public void SetInput(Mat blob, string name = "")
Parameters
Type Name Description
Mat blob

new blob.
System.String name

descriptor of the updating layer output blob.
Remarks

connect(String, String) to know format of the descriptor. If updating blob is not empty then @p blob must have the same shape, because network reshaping is not implemented yet.

| Improve this Doc View Source

SetInputsNames(IEnumerable<String>)

Sets outputs names of the network input pseudo layer.

Declaration
public void SetInputsNames(IEnumerable<string> inputBlobNames)
Parameters
Type Name Description
IEnumerable<System.String> inputBlobNames
Remarks
  • Each net always has special own the network input pseudo layer with id=0.
  • This layer stores the user blobs only and don't make any computations.
  • In fact, this layer provides the only way to pass user data into the network.
  • As any other layer, this layer can label its outputs and this function provides an easy way to do this.
| Improve this Doc View Source

SetPreferableBackend(Backend)

Ask network to use specific computation backend where it supported.

Declaration
public void SetPreferableBackend(Backend backendId)
Parameters
Type Name Description
Backend backendId

backend identifier.
| Improve this Doc View Source

SetPreferableTarget(Target)

Ask network to make computations on specific target device.

Declaration
public void SetPreferableTarget(Target targetId)
Parameters
Type Name Description
Target targetId

target identifier.

Implements