Mailman 3 July 2013 - scikit-image

GLCM with scikit-image
by elyraz＠mail.com 08 Jul '13

08 Jul '13

Hi all, I was using until now Matlab and its about time for me to move to scikit-image as it provide me more flexibility and a lot of benefits. Normally, I used Matlab to calculate the properties of gray-level co-occurrence matrix as shown in this link ( http://www.mathworks.co.uk/help/images/ref/graycoprops.html) mainly with this simple script: clc; Img = imread('C:\Users\dell\Desktop\ImgTemp\python.jpg'); I=rgb2gray(Img); % Photo downloaded from “http://i425.photobucket.com/albums/pp337/jewarmy/python.jpg” GLCM2 = graycomatrix(I); allst = graycoprops(GLCM2,'all'); contrastInfo = allst.Contrast; display(contrastInfo) energyInfo = allst.Energy; display(energyInfo) homogeneityInfo = allst.Homogeneity; display(homogeneityInfo) correlationInfo = allst.Correlation; display(correlationInfo) With the following output: contrastInfo = 0.2516 energyInfo = 0.1094 homogeneityInfo = 0.8959 correlationInfo = 0.9672 While I was trying to do it with scikit-image using this script: import numpy as np from skimage.io import imread from skimage.feature import greycomatrix, greycoprops image=imread('C:/Users/dell/Desktop/ImgTemp/python.jpg', as_grey=True) g = greycomatrix(image, [0, 1], [0, np.pi/2], levels=256) contrast = greycoprops(g, 'contrast') print('contrast is: ', contrast) energy = greycoprops(g, 'energy') print('energy is: ', energy) homogeneity = greycoprops(g, 'homogeneity') print('homogeneity is: ', homogeneity) correlation = greycoprops(g, 'correlation') print('correlation is: ', correlation) dissimilarity = greycoprops(g, 'dissimilarity') print('dissimilarity is: ', dissimilarity) ASM = greycoprops(g, 'ASM') print('ASM is: ', ASM) I get these results: contrast is: [[0 0] [0 0]] energy is: [[ 40007.37212065 40007.37212065] [ 38525.88698525 38017.06358992]] homogeneity is: [[ 165440. 165440.] [ 165088. 164970.]] correlation is: [[ 1. 1.] [ 1. 1.]] dissimilarity is: [[0 0] [0 0]] ASM is: [[1600589824 1600589824] [1484243968 1445297124]] I do not understand why there are differences and for sure I miss something. Can someone please explain me what is wrong (I am using python 3.2 and scikit-image 0.8.3). Thanks a lot in advance.

3 2

Superpixel Region Growing
by Brickle Macho 05 Jul '13

05 Jul '13

On 29/06/13 11:16 PM, StÃ¯Â¿Â½fan van der Walt wrote: > Hi Michael > > On Thu, Jun 20, 2013 at 5:37 PM, Brickle Macho <bricklemacho(a)gmail.com> wrote: >> I over segment an image using a superpixel algorithm. I region grow >> using the superpixels to end up with a segmented image, a label-image. >> I overlay the label boundaries using mark_boundaries(). > I'd be interested to hear more about your approach. We're also > working on some segmentation for gsoc, and I'd be very interested in a > region growing PR. > > StÃ¯Â¿Â½fan It not terribly exciting, perhaps calling it region growing is stretching the truth a little, not sure if it is worthy of a PR. I just traversing the superpixels using a connectivity graph (see http://peekaboo-vision.blogspot.com.au/2011/08/region-connectivity-graphs-i…) and combing neighbouring superpixels based on a decision function. Here is a little more detail on what I am doing. First I over segment the image using either slic, quickshift or felzenswalb. I am also looking at SEEDS (http://www.vision.ee.ethz.ch/~boxavier/seeds/) using the C++ implementation provided. I then visit each superpixel and determine if it needs to be combined with one or more of it neighbours. For each segment I compute a set of features and then compare each superpixel pixel with it neighbour and combine based on similarity between one or more of the features. I relabel similar superpixels to the same label, effectively "growing" superpixels. This is what I am refer to as "superpixel region growing". For me the interesting part is how do I combine the superpixels and what is a good number of superpixels to seed the process. Using various features/properties include texture, moments, orieintations, various feature histograms (normalised) etc. I plan to learn how to combine the superpixels. I also have a corresponding depth values for the image which allows me to use depth features such as average depth, surface normals, local depth patterns. I like the superpixel algorithms because of the boundary preserving property of most superpixel algorithms. Intuitively it also feel like I have the complexity of the image form MxN pixels to K superpixels, where K < MxN. In most cases traversing the "simple" graph is quicker than traversing each pixel in the image. Although in the extreme case where K approaches MxN this is not the case. Michael. --

1 0

1st and 2st order statistical texture features of an image
by Dan 04 Jul '13

04 Jul '13

Hi, I wish to perform first (histogram based mean, stdev, smoothness, skewness, uniformity and entropy) and second order (GLCM based contrast, correlation, energy, homogeneity) statistical texture features of an image. is it possible in scikit-image? If so a small script will be a huge help. Thanks.

2 2

Re: Weird - 'if' body not getting executed in spite of the expression being True
by Ankit Agrawal 03 Jul '13

03 Jul '13

Hi Chintak, Can you post what the problem was? Thanks. Regards, Ankit Agrawal, Communication and Signal Processing, IIT Bombay. On Wed, Jul 3, 2013 at 10:33 PM, Chintak Sheth <chintaksheth(a)gmail.com>wrote: > Oh boy! Thanks a ton Neil! > > Chintak > > -- > You received this message because you are subscribed to the Google Groups > "scikit-image" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to scikit-image+unsubscribe(a)googlegroups.com. > For more options, visit https://groups.google.com/groups/opt_out. > > >

1 0

Re: Weird - 'if' body not getting executed in spite of the expression being True
by Chintak Sheth 03 Jul '13

03 Jul '13

Oh boy! Thanks a ton Neil! Chintak

1 0

Weird - 'if' body not getting executed in spite of the expression being True
by Chintak Sheth 03 Jul '13

03 Jul '13

Hi everyone, It's feels stupid even as I'm typing this mail out. But as the subject states, the 'if' body is simply not getting executed. I've tried using pdb++ to debug and every time I step in after seeing that the expression should be True, the interpreter simply jumps the code block and resumes from the enclosing loop. I think it may be something funny with the indentation or so. Though PEP8 doesn't throw any warning or errors. I've retyped the statement again and again and I'm running out of ideas now. If anyone has time then please give it a look, https://github.com/chintak/scikit-image/blob/inpaint/skimage/inpaint/fmm.py. The if at line number 51 in fast_marching_method function. Thanks, Chintak

1 0

Re: segmentation fault with the viewer
by Chintak Sheth 03 Jul '13

03 Jul '13

Hi, I'm not sure what the term blocking means here but the test script seems to work fine on my system. Chintak

1 0

Re: Cython guidelines
by Andreas Mueller 03 Jul '13

03 Jul '13

On 07/02/2013 01:45 AM, Juan Nunez-Iglesias wrote: > Hi Andy, > > On Mon, Jul 1, 2013 at 5:13 PM, Andreas Mueller > <amueller(a)ais.uni-bonn.de <mailto:amueller@ais.uni-bonn.de>> wrote: > > Turns out, I didn't. That explains why they are slower than C. Not > sure why I did that... > > > Not really: > http://jakevdp.github.io/blog/2012/08/08/memoryview-benchmarks/ > > Cython + pointers is just as fast as memoryviews. And on SLIC at > least, you made liberal use of pointers. =) I just tried my hand at > memoryviews at Stéfan's request and it's quite a bit slower currently. > I did use pointers? Well as far as I remember, I tried to make it fast. So yeah, pointers make sense. Not sure what is wrong, then ... anyhow, speed is acceptable, I guess.. Thanks for giving it a shot!

2 1

Re: segmentation fault with the viewer
by Emmanuelle Gouillart 02 Jul '13

02 Jul '13

> Are you on master? Specifically, a version later than yesterday morning? The > viewer-linking PR that was merged yesterday had a number of fixes for PySide. I > don't see any issue on PyQt at the moment. Yes, I'm on master (updated this morning, and I still get the same problem). Maybe I have some libraries that are too old, I'll check on a machine with a more recent Ubuntu when I get the occasion. Thanks for your help, anyway! Emmanuelle

2 1

Re: Cython guidelines
by Andreas Mueller 02 Jul '13

02 Jul '13

On 05/29/2013 12:37 PM, Stéfan van der Walt wrote: > Hi everyone > > Two Cython related suggestions: > I just read your (admittedly quite old) mail just now and though "hey didn't I use memory views for the segmentation algorithms?" Turns out, I didn't. That explains why they are slower than C. Not sure why I did that... One day, I'll update them, I'm sure ;) Cheers, Andy

2 1