MCPcopy Create free account
hub / github.com/davisking/dlib / bind_image_classes3

Function bind_image_classes3

tools/python/src/image3.cpp:376–878  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

374// ----------------------------------------------------------------------------------------
375
376void bind_image_classes3(py::module& m)
377{
378 const char* docs;
379
380 docs =
381"requires \n\
382 - thresh > 0 \n\
383ensures \n\
384 - Converts an image to a target pixel type. dtype must be a string containing one of the following: \n\
385 uint8, int8, uint16, int16, uint32, int32, uint64, int64, float32, float, float64, double, or rgb_pixel \n\
386 \n\
387 The contents of img will be scaled to fit the dynamic range of the target \n\
388 pixel type. The thresh parameter is used to filter source pixel values which \n\
389 are outliers. These outliers will saturate at the edge of the destination \n\
390 image's dynamic range. \n\
391 - Specifically, for all valid r and c: \n\
392 - We scale img[r][c] into the dynamic range of the target pixel type. This \n\
393 is done using the mean and standard deviation of img. Call the mean M and \n\
394 the standard deviation D. Then the scaling from source to destination is \n\
395 performed using the following mapping: \n\
396 let SRC_UPPER = min(M + thresh*D, max(img)) \n\
397 let SRC_LOWER = max(M - thresh*D, min(img)) \n\
398 let DEST_UPPER = max value possible for the selected dtype. \n\
399 let DEST_LOWER = min value possible for the selected dtype. \n\
400 \n\
401 MAPPING: [SRC_LOWER, SRC_UPPER] -> [DEST_LOWER, DEST_UPPER] \n\
402 \n\
403 Where this mapping is a linear mapping of values from the left range \n\
404 into the right range of values. Source pixel values outside the left \n\
405 range are modified to be at the appropriate end of the range.";
406 /*!
407 requires
408 - thresh > 0
409 ensures
410 - Converts an image to a target pixel type. dtype must be a string containing one of the following:
411 uint8, int8, uint16, int16, uint32, int32, uint64, int64, float32, float, float64, double, or rgb_pixel
412
413 The contents of img will be scaled to fit the dynamic range of the target
414 pixel type. The thresh parameter is used to filter source pixel values which
415 are outliers. These outliers will saturate at the edge of the destination
416 image's dynamic range.
417 - Specifically, for all valid r and c:
418 - We scale img[r][c] into the dynamic range of the target pixel type. This
419 is done using the mean and standard deviation of img. Call the mean M and
420 the standard deviation D. Then the scaling from source to destination is
421 performed using the following mapping:
422 let SRC_UPPER = min(M + thresh*D, max(img))
423 let SRC_LOWER = max(M - thresh*D, min(img))
424 let DEST_UPPER = max value possible for the selected dtype.
425 let DEST_LOWER = min value possible for the selected dtype.
426
427 MAPPING: [SRC_LOWER, SRC_UPPER] -> [DEST_LOWER, DEST_UPPER]
428
429 Where this mapping is a linear mapping of values from the left range
430 into the right range of values. Source pixel values outside the left
431 range are modified to be at the appropriate end of the range.
432 !*/
433 m.def("convert_image_scaled", convert_image_scaled<uint8_t>, py::arg("img"), py::arg("dtype"), py::arg("thresh")=4);

Callers 1

PYBIND11_MODULEFunction · 0.85

Calls 1

argClass · 0.50

Tested by

no test coverage detected