fjean / pymeanshift Goto Github PK

Hi, is there any formula which allow me to get the best value for hs, hr and m?
I am doing tongue image segmentation.

How to get the density of each region as an output ?

Is there a way to get an array of segments (instead just segmented image)? 
Looks like my request could be satisfied by c++ methods GetBoundaries and 
GetRegions. But how can I access them via python?

Original issue reported on code.google.com by [email protected] on 8 May 2013 at 10:46

The example returns the number of regions (number_regions).
How do I return the list of these regions to process one by one?

What steps will reproduce the problem?
1. try the sample example in the project home page

What is the expected output? What do you see instead?

The function should return the segmented image.

when executing the simple example

import cv
from numpy import *
import pymeanshift as pms

name = "path to image"
im = cv.LoadImageM(name)
(segmentedImage, labelsImage, numberRegions) = pms.segmentMeanShift(im)


The following exception occurs

Traceback (most recent call last):
  File "meanshift.py", line 9, in <module>
    (segmentedImage, labelsImage, numberRegions) = pms.segmentMeanShift(im)
  File "/usr/local/lib/python2.7/dist-packages/pymeanshift.py", line 151, in segmentMeanShift
    nbRegion = _segmentMeanShiftImpl(inputImageMat, colorImageMat, labelImageMat, sigmaS, sigmaR, minRegion, speedUpLevel)
TypeError: %%typemap: could not convert input argument to an CvMat

i have printed the type of cvmat in line 151 and it was cv2.cv.mat could that 
be the problem ?

What version of the product are you using? On what operating system?
opencv : opencv 2.3.1
os : ubuntu 11.10


Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 1 May 2012 at 11:56

Hello!
I recently updated my anaconda and it upgraded numpy version to 1.8.1.
and I got this error when I call "_pymeanshift.segment(image, spatial_radius, range_radius, min_density, speedup_level)" in pymeanshift.py to use "segment" function

Error: NotImplementedError: PyArray_FromDims: use PyArray_SimpleNew.

Is this due to the depency of numpy?

thank you!

Traceback (most recent call last):
File "h.py", line 15, in
(segmented_image, labels_image, number_regions) = pms.segment(origin_im, spatial_radius=6, range_radius=4.5, min_density=50)
AttributeError: module 'pymeanshift' has no attribute 'segment'

When bash setup.py build it will failed:
`` raise RuntimeError("Error: Failed to import Numpy ({}). Please check that Numpy is installed on your system.".format(exc))'Change it topython setup.py build` and everything goes well

import _pymeanshift

ModuleNotFoundError: No module named '_pymeanshift'

I Tried the code on my image but i got this error!

Hi
Using this great tool for a long time but recently with new win 10 OS setup, I can’t figure problem with Visual C++ 2008 express Prerequisite, wasn’t able to find the right pieces. installation wiki has wrong or 404 link. It can be fixed? Please

Only spatial radius hs and range radius hr are talked in the original paper, but it seems the min_density in the pymeanshift package does exert influence on the segmentation results. I am wondering what is the role of this parameter and why it is not mentioned in the original paper? The spatial radius and range radius used here is the same described in the original paper?

Thanks any guys providing any insights!

I have a doubt regarding the returned segmented image. Is it in the LUV colorspace or does it return segmented image in the RGB colorspace ?

I applied the code on my image and i wanted to know the locations of each cluster to apply some operations on some clusters , where could i get them ?

Hi
There is a way to define smallest segment size ?
as i play with parameters it will produce a few segment with very small size that i am unable to merge with a neighbor segment.
thanks for all

import _pymeanshift
ModuleNotFoundError: No module named '_pymeanshift'

Hi
may i ask how i can merge/remove regions?
there is a few micro region produced over the image which i want to remove or merge.
thanks