python

Python bindings for libresdet.

The files in this directory include full bindings for the resdet API as well as additional modules supporting NumPy and Pillow data types.

Use of these bindings requires that libresdet be built as a shared library. This can be done from the resdet C sources using ./configure --enable-shared, make lib, and optionally make install-lib if installation of the library is desired.

Modules:

• resdet - The native Python interfaces for resdet.
• libresdet_api - Contains the raw ctypes mappings, this need not be used directly.
• resdet_numpy - Optional module which includes support for numpy ndarrays.
• resdet_PIL - Optional module which includes support for PIL images.

For detection with resdetect and Analysis.analyze_image, image data is accepted as any of the following types:

• A raw ctypes.POINTER(ctypes.c_float).
• A list of floats.
• An array created with type code "f".
• A numpy.ndarray with the resdet_numpy module.
• A PIL.image with the resdet_PIL module.

Examples

These examples assume image_path has been set to the path of an image. Paths may be strings or PathLike objects.

---

Run resdet and print each detected width and height:

from resdet import resdetect

resolutions = resdetect(image_path)

print("widths:")
for resolution in resolutions["widths"]:
    print(f"{resolution.index} {resolution.confidence}")

print("heights")
for resolution in resolutions["heights"]:
    print(f"{resolution.index} {resolution.confidence}")

resdet's detection parameters can be controlled by calling resdetect or the Analysis constructor with parameters = { "threshold": the_threshold, "range": the_range, "compression_filter": the_compression_filter } A detection method can be provided with method = the_method. Methods can be obtained as a list using resdetect.methods().

---

Run resdet on a NumPy ndarray:

from resdet_numpy import resdetect

resolutions = resdetect(myndarray)

---

Run resdet on a Pillow image:

from resdet_PIL import resdetect

image = PIL.Image.open(image_path)

resolutions = resdetect(image)

---

Use resdet's image reading API together with its sequential analysis API. Note that this example is equivalent to calling resdetect(image_path):

from resdet import ImageBuffer, Image, Analysis

imagebuf = ImageBuffer()
image = Image(image_path, buffer = imagebuf)
analysis = Analysis(image.width, image.height)

while image.read_image_frame(imagebuf):
    analysis.analyze_image(imagebuf)

resolutions = analysis.analysis_results()

These classes are also context managers, so the equivalent code using with is:

with ImageBuffer() as imagebuf:
    with Image(image_path, buffer = imagebuf) as image:
        with Analysis(image.width, image.height) as analysis:
            while image.read_image_frame(imagebuf):
              analysis.analyze_image(imagebuf)
            resolutions = analysis.analysis_results()