I want to do some image OCR with PyTesseract, and I've seen that OpenCV's erode and dilate functions are very useful for noise removal pre-processing.
Since PyTesseract already requires PIL/Pillow, I'd like to do the noise removal in PIL, rather than get another library. Is there an equivalent to erode/dilate in PIL? (My research seems to suggest that MaxFilter and MinFilter could be used this way, but it's not fully clear to me if that's really true.)
Thanks!
The best option is to use OpenCV python bindings. However, if you want to use PIL/Pillow, there is the ImageFilter Module:
http://pillow.readthedocs.io/en/3.1.x/reference/ImageFilter.html
dilation_img = src_img.filter(ImageFilter.MaxFilter(3))
erosion_img = src_img.filter(ImageFilter.MinFilter(3))
The number 3 in the example is the mask size;
Related
I've already made a function to cut out the image, and the part I cut out has a black background. I'm trying to make the black part transparent, then I can generate an image sequence that I can create a video with. I've tried converting the image to a double and then replacing the 0 values with NaN:
J = imread('imgExample.jpg');
J2 = im2double(J);
J2(J2 == 0) = NaN;
imwrite(J2, 'newImg.jpg');
but when I convert it into a video, it doesn't seem to stay. Is there any way to get the black part of the image to be transparent?
From clarifications in comments, you are trying to create a video format that supports alpha transparency using matlab.
In general this seems impossible using matlab alone (at least in matlab 2013 which is the version I use). If you'd like to check if the newest matlab supports videos with alpha transparency, type doc videowriter and have a look at the available formats. If you see anything with transparency options there, take it from there. But the most I see on mine is 24bit RGB videos (i.e. three channels, no transparency).
So matlab does not have the ability to produce native .avi video with alpha transparency.
However, note that this is a very rare video format anyway, and even if you did manage to produce such a video, you would still have to find a suitable viewer which supports playing videos with transparency!
It's therefore important for you to tell us your particular use-case because it may be you're actually trying to do something much simpler (which may or may not be solvable via matlab) (i.e. a case of the XY Problem
E.g. you may be trying to create a video with transparency for the web instead, like here https://developers.google.com/web/updates/2013/07/Alpha-transparency-in-Chrome-video
If this is the case, then I would recommend you attempting the method outlined there; you can create individual .png "frames" with transparency in matlab using the imwrite function. have a look at its documentation, particularly the section about png images and the 'Alpha' property. But beyond that, you'd need an external tool to combine them into a .webm file, since matlab doesn't seem to have a tool like that (at least none that I can see at a glance; there might be a 3rd-party toolkit if you look on the web).
Hope this helps.
For a school project I've built a scanner and connected it to matlab. The scanner scans images (16-by-16 pixels) of handwritten digits from 0 to 9. I'm using a principal component analysis in order to classify the scans. Due to the low accuracy of the scanner, I need to preprocess the scans first, before I can actually send them through the recognition machine.
One of these preprocessing-steps is to thicken the lines. So far, I've used a pretty simple averageing filter for this: H = ones(3, 3) ./ 9. This bears the problem, that the circular gap of the digits 8 and 9 is likely to be "closed". I enclose a picture of all my preprocessing-steps, where the problem is visible: the image with the caption "threshholded" still shows the gap, but it disappeared after the thickening step.
My question is: Do you know a better filter for this "thickening"-step, which would not erase the gap? Or do you have an idea for a filter which could be applied after the thickening to produced the desired result? Any other suggestions or hints are also greatly appreciated.
I=imread('numberreco.png');
subplot(1,2,1),imshow(I)
I=rgb2gray(I);
BW=~im2bw(I,graythresh(I));
BW2 = bwmorph(BW,'thin');
I1=double(I).*BW2;
subplot(1,2,2),imshow(uint8(I1))
The gap is kept, and you can start from here...
Not a very general answer, but if you have the Image Processing Toolbox, and your system doesn't depend on having multiple grey levels, then converting to binary images and using the 'thicken' operation from bwmorph() should do exactly what you want.
Thinking a bit harder, you could also use a suitably thickened binary image as a mask to restore holes - either just elementwise multiply it with the blurred greyscale image or, for more flexibility:
invert it to form a background/holes mask
remove the background with imclearborder() to leave just the holes
optionally dilate the mask
use as a logical index to clear the 'hole' areas of the blurred/brightened greyscale image.
Even without the morphological steps you can use a mask to artificially reintroduce the original holes later, e.g.:
bgmask = (thresholdedimage == 0); % assuming 0 == background
holes = imclearborder(bgmask);
... % other processing steps
brightenedimage(holes) = 0; % punch holes in updated image
I have 42 variables and I have calculated the correlation matrix for them in Matlab. Now I would like to visualize it with a schemaball. Does anyone have any suggestions / experiences how this could be done in Matlab? The following pictures will explain my point better:
In the pictures each parabola between variables would mean the strength of correlation between them. The thicker the line is, the more correlation. I prefer the style of picture 1 more than the style in picture 2 where I have used different colors to highlight the strength of correlation.
Kinda finished I guess.. code can be found here at github.
Documentation is included in the file.
The yellow/magenta color (for positive/negative correlation) is configurable, as well as the fontsize of the labels and the angles at which the labels are plotted, so you can get fancy if you want and not distribute them evenly along the perimeter/group some/...
If you want to actually print these graphs or use them outside matlab, I suggest using vector formats (eg eps). It's also annoying that the text resizes when you zoom in/out, but I don't know of any way to fix that without hacking the zoom function :/
schemaball % demo
schemaball(arrayfun(#num2str,1:10,'uni',false), rand(10).^8,11,[0.1587 0.8750],[0.8333 1],2*pi*sin(linspace(0,pi/2-pi/20,10)))
schemaball(arrayfun(#num2str,1:50,'uni',false), rand(50).^50,9)
I finished and submitted my version to the FEX: schemaball and will update the link asap.
There are a some differences with Gunther Struyf's contribution:
You can return the handles to the graphic object for full manual customization
Labels are oriented to allow maximum left-to-rigth readability
The figure stretches to fit labels in, leaving the axes unchanged
Syntax requires only correlations matrix (but allows optional inputs)
Optimized for performance.
Follow examples of demo, custom labels and creative customization.
Note: the first figure was exported with saveas(), all others with export_fig.
schemaball
x = rand(10).^3;
x(:,3) = 1.3*mean(x,2);
schemaball(x, {'Hi','how','is','your','day?', 'Do','you','like','schemaballs?','NO!!'})
h = schemaball;
set(h.l(~isnan(h.l)), 'LineWidth',1.2)
set(h.s, 'MarkerEdgeColor','red','LineWidth',2,'SizeData',100)
set(h.t, 'EdgeColor','white','LineWidth',1)
The default colormap:
To improve on screen rendering you can launch MATLAB with the experimental -hgVersion 2 switch which produces anti/aliased graphics by default now (source: HG2 update | Undocumented Matlab). However, if you try to save the figure, the file will have the usual old anti-aliased rendering, so here's a printscreen image of Gunther's schemaball:
Important update:
You can do this in Matlab now with the FileExchange submission:
http://www.mathworks.com/matlabcentral/fileexchange/48576-circulargraph
There is an exmample by Matlab in here:
http://uk.mathworks.com/examples/matlab/3859-circular-graph-examples
Which gives this kind of beautiful plots:
Coincidentally, Cleve Moler (MathWorks Chief Mathematician) showed an example of just this sort of plot on his most recent blog post (not nearly as beautiful as the ones in your example, and the connecting lines are straight rather than parabolic, but it looks functional). Unfortunately he didn't include the code directly, but if you leave him a comment on the post he's usually very willing to share things.
What might be even nicer for you is that he also applies (and this time includes) code to permute the rows/columns of the array in order to maximize the spatial proximity of highly connected nodes, rather than randomly ordering them around the circumference. You end up with a 'crescent'-shaped envelope of connecting lines, with the thick bit of the crescent representing the most highly connected nodes.
Unfortunately however, I suspect that if you need to enhance his code to get the very narrow, high-resolution lines in your example plots, then MATLAB's currently non-anti-aliased graphics aren't quite up to it yet.
I've recently been experimenting with MATLAB data and the D3 visualization library for similar graphs - there are several related types of circular visualizations you may be interested in and many of them are interactive. Another helpful, well-baked, and freely available option is Circos which is probably responsible for most of the prettier versions of these graphs you've seen in popular press.
I am looking to export my MATLAB plot as a high quality figure. Specifically, I would like to save it as a vector based file format such as EPS or SVG.
I have tried print and saveas commands:
saveas(h,'myFileName','epsc2');
print('-r150','-depsc2', 'myFilename');
On all occasions this produces poor quality parts of the graph, although the axis-labels are indeed vector. Why does MatLab do some horrible rendering before putting it into an EPS?
Example of poor quality plot here:
http://users.ox.ac.uk/~pemb2372/myFileName.eps
Edit:
It is also worth noting that if you use a Mac viewing an EPS file from Matlab, 'Preview' app may render inner graph content rasterized and poor quality, while leaving the axis and labels vectorized and high quality. This is very misleading but when you open said EPS file in, for example, Inkscape, the quality is actually vector and quite high.
Edit 2:
My university hosting account has expired, so you can no longer view the figure. Suffice it to say that it showed a poor quality raster-style plot within high quality beautiful axis lines, ticks and labels.
I thought I would share the issue I had, and how I overcame it...
I was getting terrible results because I had the wrong renderer set to default. In my startup.m, I had the zbuffer renderer enabled. This is an example eps output.
I made that eps output with: print(gcf,'-depsc2','filename.eps'). This eps is so OBVIOUSLY rasterised. It makes me angry at matlab. Then, I had a brainwave - perhaps my default renderer zbuffer is interfering with the image save process. So, adding the line:
set(gcf,'renderer','painters')
and running the print command as before, here is the output:
Note that I just took screenshots of the eps output files at 100%. And I can confirm the second image is actually vector. Here is a good question/explanation on choosing Renderers in MATLAB.
Matlab can export to pdf with better quality than EPS, but with its own caveats of setting decent margins and font sizes.
edit:
Examples are similar to the EPS case as explained in the help of e.g. print:
saveas(gcf,'filename.pdf')
or
print('-dpdf','filename.pdf')
You might also want to take a look at the PaperSize, PaperPosition and PaperUnits properties of your figure (by means of the set and get functions).
edit: Another option is to use one of the functions available on FileExchange such as the ones mentioned by #user664303 below. My personal favorite for use with LaTeX is matlab2tikz for which the latest version can be gotten from GitHub. Together with the external library of TikZ, I think this delivers some of the most nicest graphs around.
Probably it's also best to mention that I have been actively involved in the matlab2tikz project since 2012.
The export_fig function on the MATLAB file exchange is a reasonably reliable way of accurately exporting figures to eps and pdf (as well as bitmap formats) in MATLAB.
The plot2svg function, also from the file exchange, allows you to export in svg format. It provides some additional benefits, such as being able to export translucent patch objects in vector format.
A comparison of exporting methods is given in this blog post.
I always acquire the final plots (those which are supposed to be inserted into papers and publications) by matplotlib library of python.
You can bet on the amazing quality of the generated plots, both .pdf and .eps formats.
I'm trying to do top hat filtering in MATLAB. The imtophat function looks promising, but I have no idea how to use it. I don't have a lot of work with MATLAB before. I am trying to look find basically small spots several pixels wide that are local maxima in my 2 dimensional array.
I think you have more problem undertanding how to use STREL, than IMTOPHAT. The later can be described as simple threshold but per structural element, not the whole image.
Here is another good examples of using STREL and IMTOPHAT:
http://www.mathworks.com/matlabcentral/fx_files/2573/1/content/html/R14_MicroarrayImage_CaseStudy.html
This series of posts on Steve Eddins blog might be useful for you:
http://blogs.mathworks.com/steve/category/dilation-algorithms/
tophat is basically an "opening" procedure followed by a subtraction of the result from the original image. the best and most helpful explanation of opening I've found here:
http://homepages.inf.ed.ac.uk/rbf/HIPR2/morops.htm
"The effect of opening can be quite easily visualized. Imagine taking
the structuring element and sliding it around inside each foreground
region, without changing its orientation. All pixels which can be
covered by the structuring element with the structuring element being
entirely within the foreground region will be preserved. However, all
foreground pixels which cannot be reached by the structuring element
without parts of it moving out of the foreground region will be eroded
away."
The documentation on imtophat has an example .. did you try it? The following images are from the MATLAB documentation.
Code
I = imread('rice.png');
imshow(I)
se = strel('disk',12);
J = imtophat(I,se);
figure, imshow(J,[])
Original
(image source: mathworks.com)
Top Hat with a disk structuring element
(image source: mathworks.com)