Semantic Segmentation(语义分割)#

fastdeploy.vision.segmentation.PaddleSegPreprocessor#

class fastdeploy.vision.segmentation.PaddleSegPreprocessor(config_file)[source]#

Create a preprocessor for PaddleSegModel from configuration file

Parameters

config_file – (str)Path of configuration file, e.g ppliteseg/deploy.yaml

disable_normalize()[source]#

This function will disable normalize in preprocessing step.

disable_permute()[source]#

This function will disable hwc2chw in preprocessing step.

property is_vertical_screen#

Atrribute of PP-HumanSeg model. Stating Whether the input image is vertical image(height > width), default value is False

Returns

value of is_vertical_screen(bool)

run(input_ims)[source]#

Preprocess input images for PaddleSegModel

Parameters

input_ims – (list of numpy.ndarray)The input image

Returns

list of FDTensor

fastdeploy.vision.segmentation.PaddleSegModel#

class fastdeploy.vision.segmentation.PaddleSegModel(model_file, params_file, config_file, runtime_option=None, model_format=<ModelFormat.PADDLE: 1>)[source]#

Load a image segmentation model exported by PaddleSeg.

Parameters

  • model_file – (str)Path of model file, e.g unet/model.pdmodel

  • params_file – (str)Path of parameters file, e.g unet/model.pdiparams, if the model_fomat is ModelFormat.ONNX, this param will be ignored, can be set as empty string

  • config_file – (str) Path of configuration file for deploy, e.g unet/deploy.yml

  • runtime_option – (fastdeploy.RuntimeOption)RuntimeOption for inference this model, if it’s None, will use the default backend on CPU

  • model_format – (fastdeploy.ModelForamt)Model format of the loaded model

batch_predict(image_list)[source]#

Predict the segmentation results for a batch of input images

Parameters

image_list – (list of numpy.ndarray) The input image list, each element is a 3-D array with layout HWC, BGR format

Returns

list of SegmentationResult

clone()[source]#

Clone PaddleSegModel object

Returns

a new PaddleSegModel object

get_profile_time()#

Get profile time of Runtime after the profile process is done.

property postprocessor#

Get PaddleSegPostprocessor object of the loaded model

Returns

PaddleSegPostprocessor

predict(image)[source]#

Predict the segmentation result for an input image

Parameters

im – (numpy.ndarray)The input image data, 3-D array with layout HWC, BGR format

Returns

SegmentationResult

property preprocessor#

Get PaddleSegPreprocessor object of the loaded model

Returns

PaddleSegPreprocessor

fastdeploy.vision.segmentation.PaddleSegPostprocessor#

class fastdeploy.vision.segmentation.PaddleSegPostprocessor(config_file)[source]#

Create a postprocessor for PaddleSegModel from configuration file

Parameters

config_file – (str)Path of configuration file, e.g ppliteseg/deploy.yaml

property apply_softmax#

Atrribute of PaddleSeg model. Stating Whether applying softmax operator in the postprocess, default value is False

Returns

value of apply_softmax(bool)

run(runtime_results, imgs_info)[source]#

Postprocess the runtime results for PaddleSegModel

Parameters

  • runtime_results – (list of FDTensor)The output FDTensor results from runtime

  • imgs_info – The original input images shape info map, key is “shape_info”, value is [[image_height, image_width]]

Returns

list of SegmentationResult(If the runtime_results is predict by batched samples, the length of this list equals to the batch size)

property store_score_map#

Atrribute of PaddleSeg model. Stating Whether storing score map in the SegmentationResult, default value is False

Returns

value of store_score_map(bool)