MCPcopy Create free account
hub / github.com/CVCUDA/CV-CUDA / ExportOpSIFT

Function ExportOpSIFT

python/mod_cvcuda/operators/OpSIFT.cpp:207–274  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

205} // namespace
206
207void ExportOpSIFT(py::module &m)
208{
209 using namespace pybind11::literals;
210
211 m.def("sift", &SIFT, "src"_a, "max_features"_a = 0, "num_octave_layers"_a = 3, "contrast_threshold"_a = 0.03f,
212 "edge_threshold"_a = 10.f, "init_sigma"_a = 1.6f, "flags"_a = NVCV_SIFT_USE_EXPANDED_INPUT, py::kw_only(),
213 "stream"_a = nullptr, R"pbdoc(
214
215 Executes the SIFT operation on the given cuda stream.
216
217 See also:
218 Refer to the CV-CUDA C API reference for the SIFT operator
219 for more details and usage examples.
220
221 Args:
222 src (cvcuda.Tensor): Input tensor to extract features and compute descriptors from.
223 max_features (Number, optional): Maximum number of features to be extracted, default is 5% of total
224 pixels at a minimum of 1.
225 num_octave_layers (Number, optional): Number of octave layers, default is 3.
226 contrast_threshold (Number, optional): Contrast threshold, default is 0.03.
227 edge_threshold (Number, optional): Edge threshold, default is 10.0.
228 init_sigma (Number, optional): Initial sigma, default is 1.6.
229 flags (cvcuda.SIFT, optional): Flag to whether to expand the input or not, default is
230 cvcuda.SIFT.USE_EXPANDED_INPUT.
231 stream (cvcuda.Stream, optional): CUDA Stream on which to perform the operation.
232
233 Returns:
234 Tuple[cvcuda.Tensor, cvcuda.Tensor, cvcuda.Tensor, cvcuda.Tensor]: A tuple with feature coordinates, metadata, descriptors and
235 number of features.
236
237 Caution:
238 Restrictions to several arguments may apply. Check the C
239 API references of the CV-CUDA operator.
240 )pbdoc");
241
242 m.def("sift_into", &SIFTInto, "feat_coords"_a, "feat_metadata"_a, "feat_descriptors"_a, "num_features"_a, "src"_a,
243 "num_octave_layers"_a = 3, "contrast_threshold"_a = 0.03f, "edge_threshold"_a = 10.f, "init_sigma"_a = 1.6f,
244 "flags"_a = NVCV_SIFT_USE_EXPANDED_INPUT, py::kw_only(), "stream"_a = nullptr, R"pbdoc(
245
246 Executes the SIFT operation on the given cuda stream.
247
248 See also:
249 Refer to the CV-CUDA C API reference for the SIFT operator
250 for more details and usage examples.
251
252 Args:
253 feat_coords (cvcuda.Tensor): Output tensor with feature coordinates.
254 feat_metadata (cvcuda.Tensor): Output tensor with feature metadata.
255 feat_descriptors (cvcuda.Tensor): Output tensor with feature descriptors.
256 num_features (cvcuda.Tensor): Output tensor with number of features.
257 src (cvcuda.Tensor): Input tensor to extract features and compute descriptors from.
258 num_octave_layers (Number, optional): Number of octave layers, default is 3.
259 contrast_threshold (Number, optional): Contrast threshold, default is 0.03.
260 edge_threshold (Number, optional): Edge threshold, default is 10.0.
261 init_sigma (Number, optional): Initial sigma, default is 1.6.
262 flags (cvcuda.SIFT, optional): Flag to whether to expand the input or not, default is
263 cvcuda.SIFT.USE_EXPANDED_INPUT.
264 stream (cvcuda.Stream, optional): CUDA Stream on which to perform the operation.

Callers 1

PYBIND11_MODULEFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected