Nodes

The most important concept of the PyCVF is the concept of node.

A node simply represent the samething as a node in “dataflow oriented processing”.

Thus, a node is where the computation are effectively occuring.

In a certain extent they may exists all kind of node, but to provide an easy syntax all nodes are assumed at first to

to have only one input, and only one output. We network realized

by connecting inputs and outputs together is called the dataflow.

Abstract Database Type

class pycvf.nodes.SPECS.ref.Node(*args, **kwargs)

Nodes or Nodes are means to transform information.

Abstractly each node is assumed to have :

1 Input 1 Output 1 Set of parametes

This interface constraint is actually not really a constraint since not that much drastic, since user may provide different outputs by outputing tuples. And may get different inputs by inputing tuples our using the parameters as inputs.

Note that there exists convenients decorators pycvf_node_function and pycvf_node_class for transforming any function or class into a PyCVF model.

Existing Nodes

pycvf.nodes.interactive

pycvf.nodes.interactive.select_node()

This functions allows you to select a node and instantiate it with its default parameters.

pycvf.nodes.naive

class pycvf.nodes.naive.Node(*args, **kwargs)

The simplest model one may imagine. It is just the identity function.

init_featurefilter()

pycvf.nodes.combined

class pycvf.nodes.combined.Combiner(*models)

This allows you to combine result of different models into a single 1.

With this node you may define one descriptor that

contain one part for sound and one part for video.

@param models different models passed as arguments either as string expressions either as already instantiated models

get_submodels_status()

init()

on_model_destroy(model)

process(x)

set_curdb(cdb)

set_model_node(model)

pycvf.nodes.exploded_transform

class pycvf.nodes.exploded_transform.Node(*args, **kwargs)

init_model(model, structure=None, modelelementpath=-1, addressed=False, recomposed=True, subsample=<function <lambda> at 0x80555f0>)

input_datatype(x)

on_destroy()

output_datatype(x)

set_curdb_(cdb)

class pycvf.nodes.reference.Reference(node='src')

This node is use to make reference to other elements in the pipeline

on_model_destroy(model)

process(points)

set_model_node(model)

pycvf.nodes.histogram.histogram_alt(x, *args, **kwargs)

class pycvf.nodes.vectors.combined.Node(*args, **kwargs)

init_model(*args)

input_datatype(x)

output_datatype(x)

class pycvf.nodes.vectors.PCA.IPCAProcessor(*args, **kwargs)

basis(i)

init(odim, burnin=300, *args, **kwargs)

iprocess(v)

on_model_destroy(model)

process(v)

save()

set_model_node(model)

class pycvf.nodes.vectors.train.NodeTrainer(statmodel, filename=None, negative_data=None, with_label=None, label_op=None, *args, **kwargs)

on_model_destroy(model)

process(x)

set_model_node(model)

class pycvf.nodes.vectors.whiten.Whitener(learning_enabled=None, filename='whitening_model')

init()

on_model_destroy(model)

process(x)

set_model_node(model)

class pycvf.nodes.sequence.select.Node(*args, **kwargs)

init_model(n, *args, **kwargs)

classmethod input_datatype(x)

output_datatype(x)

pycvf.nodes.image.segment.opencv_pyrsegment