Base Class Reference

Inheritance diagram for Base:

Collaboration diagram for Base:

Public Member Functions
def	__init__ (self, theDetector, theTracker, trackFilter, Params params)
def	det_mask (self)
def	draw_layer (self, img=None, raw_detect=False, ax=None)
def	get_mask (self)
def	get_state (self)
def	measure (self, I)
def	update_params (self, name, val)

Public Attributes
	layer_mask
	layer_mask_det
	layer_state
	params

Constructor & Destructor Documentation

__init__()

def __init__	(		self,
			theDetector,
			theTracker,
			trackFilter,
		Params	params
	)

Base class for the layer segmentation approach

Build upon the the simple detector -> tracker -> trackfilter pipeline

Now building a base classs of the following process pipeline:
preprocess -> detect -> postprocess -> track -> trackfilter

where:
- preprocess: the preprocess of the input image
- Postprocess: post process of the detected layer mask
- params: need to has the field of the Params class 

Member Function Documentation

det_mask()

def det_mask

(

self

)

Make the default getting the foregound mask

To be overwritten for any specific detector type

Reimplemented in Puzzle_Residual, robot_inRange, Human_ColorSG, Base_bg, and Base_fg.

draw_layer()

def draw_layer	(		self,
			img = None,
			raw_detect = False,
			ax = None
	)

Visualize the layer result

@ param[in] img             The input image. Default is None. If not None, then will crop the layer mask area and show.
                    If None, then will only plot the binary mask
@ param[in] raw_detect      bool. Default is False. If set to true, will draw the raw detected mask without postprocessing, 
                    and will not display the trackerstate

get_mask()

def get_mask

(

self

)

get_state()

def get_state

(

self

)

measure()

def measure	(		self,
			I
	)

Set a common processing pipeline?

But different detector or tracker will generate different result name. 
e.g. the layer mask should be obtained from fg_detector.getForeGround() and bg_detector.getBackground() separetely
similar for the trackers 

Might be better off defining the pipeline separately for different subclass of segmentor? Or just make up some default?

TODO: here requires the tracker instance to have the process & getstate API.
might be better to also limit the input to some base tracker class with those APIs?

Reimplemented in Human_ColorSG_HeightInRange.

update_params()

def update_params	(		self,
			name,
			val
	)

Update a parameter

Member Data Documentation

layer_mask

layer_mask

layer_mask_det

layer_mask_det

layer_state

layer_state

params

params

---

The documentation for this class was generated from the following file:

• /home/pvela/python/surveillance/Surveillance/layers/base.py