public void Classify(
IInputArray frame,
out int classId,
out float conf)
{
classId = -1;
conf = 0;
using (InputArray iaFrame = frame.GetInputArray())
{
DnnInvoke.cveDnnClassificationModelClassify(
_ptr,
iaFrame,
ref classId,
ref conf);
}
}
/// <summary>
/// Given the input frame, create input blob, run net and return result detections.
/// </summary>
/// <param name="frame">The input image.</param>
/// <param name="classIds">Class indexes in result detection.</param>
/// <param name="confidences">A set of corresponding confidences.</param>
/// <param name="boxes">A set of bounding boxes.</param>
/// <param name="confThreshold">A threshold used to filter boxes by confidences.</param>
/// <param name="nmsThreshold">A threshold used in non maximum suppression.</param>
public void Detect(
IInputArray frame,
VectorOfInt classIds,
VectorOfFloat confidences,
VectorOfRect boxes,
float confThreshold = 0.5f,
float nmsThreshold = 0.5f)
{
using (InputArray iaFrame = frame.GetInputArray())
{
DnnInvoke.cveDnnDetectionModelDetect(
_ptr,
iaFrame,
classIds,
confidences,
boxes,
confThreshold,
nmsThreshold);
}
}
/// <summary>
/// Create model from deep learning network.
/// </summary>
/// <param name="net">DNN Network</param>
public DetectionModel(Net net)
{
_ptr = DnnInvoke.cveDnnDetectionModelCreate2(
net,
ref _model);
}
/// <summary>
/// Create Text Recognition model from deep learning network
/// </summary>
/// <param name="net">Dnn network</param>
/// <remarks>Set DecodeType and Vocabulary after constructor to initialize the decoding method.</remarks>
public TextRecognitionModel(Net net)
{
_ptr = DnnInvoke.cveDnnTextRecognitionModelCreate2(
net,
ref _model);
}
/// <summary>
/// Default constructor.
/// </summary>
public Net()
{
_ptr = DnnInvoke.cveDnnNetCreate();
}
/// <summary>
/// Ask network to make computations on specific target device.
/// </summary>
/// <param name="value">The value</param>
public void SetPreferableTarget(Target value)
{
DnnInvoke.cveNetSetPreferableTarget(_ptr, value);
}
/// <summary>
/// Constructs 4-dimensional blob (so-called batch) from image or array of images.
/// </summary>
/// <param name="image">2-dimensional multi-channel or 3-dimensional single-channel image (or array of images)</param>
public Blob(IInputArray image)
{
using (InputArray iaImage = image.GetInputArray())
_ptr = DnnInvoke.cveDnnBlobCreateFromInputArray(iaImage);
}
/// <summary>
/// Ask network to use specific computation backend where it supported.
/// </summary>
/// <param name="value">The value</param>
public void SetPreferableBackend(Backend value)
{
DnnInvoke.cveModelSetPreferableBackend(_model, value);
}
/// <summary>
/// Set flag crop for frame.
/// </summary>
/// <param name="crop">Flag which indicates whether image will be cropped after resize or not.</param>
public void SetInputCrop(bool crop)
{
DnnInvoke.cveModelSetInputCrop(_model, crop);
}
/// <summary>
/// Dump net structure, hyperparameters, backend, target and fusion to dot file
/// </summary>
/// <param name="path">Path to output file with .dot extension</param>
public void DumpToFile(String path)
{
using (CvString p = new CvString(path))
DnnInvoke.cveDnnNetDumpToFile(_ptr, p);
}
/// <summary>
/// Adds loaded layers into the <paramref name="net"/> and sets connetions between them.
/// </summary>
/// <param name="net">The net model</param>
public void PopulateNet(Net net)
{
DnnInvoke.cveDnnImporterPopulateNet(_ptr, net);
}
/// <summary>
/// Runs forward pass for the whole network.
/// </summary>
public void Forward()
{
DnnInvoke.cveDnnNetForward(_ptr);
}
/// <summary>
/// Sets the new value for the layer output blob.
/// </summary>
/// <param name="outputName">Descriptor of the updating layer output blob.</param>
/// <param name="blob">New blob</param>
public void SetBlob(String outputName, Blob blob)
{
using (CvString outputNameStr = new CvString(outputName))
DnnInvoke.cveDnnNetSetBlob(_ptr, outputNameStr, blob);
}
/// <summary>
/// Create model from deep learning network.
/// </summary>
/// <param name="net">DNN Network</param>
public KeypointsModel(Net net)
{
_ptr = DnnInvoke.cveDnnKeypointsModelCreate2(
net,
ref _model);
}
/// <summary>
/// Create model from deep learning network.
/// </summary>
/// <param name="network">DNN Network</param>
public Model(Net network)
{
_ptr = DnnInvoke.cveModelCreateFromNet(network);
_model = _ptr;
}
/// <summary>
/// Set mean value for frame.
/// </summary>
/// <param name="mean">Scalar with mean values which are subtracted from channels.</param>
public void SetInputMean(MCvScalar mean)
{
DnnInvoke.cveModelSetInputMean(_model, ref mean);
}
/// <summary>
/// Set input size for frame.
/// </summary>
/// <param name="size">New input size.</param>
/// <remarks>If shape of the new blob less than 0, then frame size not change.</remarks>
public void SetInputSize(Size size)
{
DnnInvoke.cveModelSetInputSize(_model, ref size);
}
/// <summary>
/// Sets the new value for the layer output blob.
/// </summary>
/// <param name="name">Descriptor of the updating layer output blob.</param>
/// <param name="blob">Input blob</param>
/// <param name="scaleFactor">An optional normalization scale.</param>
/// <param name="mean">An optional mean subtraction values.</param>
public void SetInput(IInputArray blob, String name = "", double scaleFactor = 1.0, MCvScalar mean = new MCvScalar())
{
using (CvString nameStr = new CvString(name))
using (InputArray iaBlob = blob.GetInputArray())
DnnInvoke.cveDnnNetSetInput(_ptr, iaBlob, nameStr, scaleFactor, ref mean);
}
/// <summary>
/// Set flag swapRB for frame.
/// </summary>
/// <param name="swapRB">Flag which indicates that swap first and last channels.</param>
public void SetInputSwapRB(bool swapRB)
{
DnnInvoke.cveModelSetInputSwapRB(_model, swapRB);
}
/// <summary>
/// Runs forward pass to compute outputs of layers listed in outBlobNames.
/// </summary>
/// <param name="outputBlobs">Contains blobs for first outputs of specified layers.</param>
/// <param name="outBlobNames">Names for layers which outputs are needed to get</param>
public void Forward(IOutputArrayOfArrays outputBlobs, String[] outBlobNames)
{
using (OutputArray oaOutputBlobs = outputBlobs.GetOutputArray())
using (VectorOfCvString vcs = new VectorOfCvString(outBlobNames))
DnnInvoke.cveDnnNetForward3(_ptr, oaOutputBlobs, vcs);
}
/// <summary>
/// Ask network to make computations on specific target device.
/// </summary>
/// <param name="value">The value</param>
public void SetPreferableTarget(Target value)
{
DnnInvoke.cveModelSetPreferableTarget(_model, value);
}
/// <summary>
/// Create model from deep learning network.
/// </summary>
/// <param name="net">DNN Network</param>
public SegmentationModel(Net net)
{
_ptr = DnnInvoke.cveDnnSegmentationModelCreate2(
net,
ref _model);
}
/// <summary>
/// Ask network to use specific computation backend where it supported.
/// </summary>
/// <param name="value">The value</param>
public void SetPreferableBackend(Backend value)
{
DnnInvoke.cveNetSetPreferableBackend(_ptr, value);
}
public void BatchFromImages(IInputArray image, int dstCn = -1)
{
using (InputArray iaImage = image.GetInputArray())
DnnInvoke.cveDnnBlobBatchFromImages(_ptr, iaImage, dstCn);
}
/// <summary>
/// Enables or disables layer fusion in the network.
/// </summary>
/// <param name="value">The value</param>
public void EnableFusion(bool value)
{
DnnInvoke.cveNetEnableFusion(_ptr, value);
}
public IntPtr GetPtr(int n = 0, int cn = 0, int row = 0, int col = 0)
{
return(DnnInvoke.cveDnnBlobGetPtr(_ptr, n, cn, row, col));
}
/// <summary>
/// Sets the new value for the layer output blob.
/// </summary>
/// <param name="name">Descriptor of the updating layer output blob.</param>
/// <param name="blob">Input blob</param>
public void SetInput(Mat blob, String name)
{
using (CvString outputNameStr = new CvString(name))
DnnInvoke.cveDnnNetSetInput(_ptr, blob, outputNameStr);
}
public Blob()
{
_ptr = DnnInvoke.cveDnnBlobCreate();
}
/// <summary>
/// Set the decoding method options for "CTC-prefix-beam-search" decode usage
/// </summary>
/// <param name="beamSize">Beam size for search</param>
/// <param name="vocPruneSize">Parameter to optimize big vocabulary search, only take top <paramref name="vocPruneSize"/> tokens in each search step, <paramref name="vocPruneSize"/> <= 0 stands for disable this prune.</param>
public void SetDecodeOptsCTCPrefixBeamSearch(int beamSize, int vocPruneSize)
{
DnnInvoke.cveDnnTextRecognitionModelSetDecodeOptsCTCPrefixBeamSearch(_ptr, beamSize, vocPruneSize);
}
/// <summary>
/// Set scalefactor value for frame.
/// </summary>
/// <param name="scale">Multiplier for frame values.</param>
public void SetInputScale(double scale)
{
DnnInvoke.cveModelSetInputScale(_model, scale);
}
