8

The sample image you posted has relatively strong perspective (it is not imaged straight from the direction of surface normal) which can cause problems with template matching techniques witch use block processing. I assume that you have to take the image with strong perspective so first thing we want to do is estimate image transformation which will remove ...


5

I hesitate to write this as an answer, but given that you're asking only for advice, I will do so. I suggest investigating techniques based on the Dual-Tree Complex Wavelet Transform (DTCWT). These have shown to be useful for generating descriptors that have good tolerance to shift, scale and rotation of the source images. It's not the classic problem in ...


5

I think that you can solve you problem in a much easier way. Considering that you are dealing with blueprints, you should not worry about edge connectivity, noise, and many other things that SIFT and SURF were built to accommodate for. Your template is a hollow shape with specific edge shapes. Thus, My recommendation is: Walk around the perimeter and find ...


4

I suspect your problem occurs due to some scaling issues. Basically you need to normalize your research image to the pattern template by subtracting the mean value of the template. And it is better calculate the ratio of correlation to the standard deviation of both images. I don't know which programming language you are using. I wrote a Matlab code for you ...


3

I have a PRN generator that I have validated with live captured signals that is available on the Mathworks Exchange site at this address and equally runs in Octave (Update: I also pasted the core of this in a code block below): https://www.mathworks.com/matlabcentral/fileexchange/14670-gps-c-a-code-generator The two tap coder is as given in the diagram in ...


3

If your template or kernel is small, then straight convolution might be the fastest approach. There's a crossover point when performing convolution in the frequency domain is faster than straight time/spatial domain convolution and it can be hardware dependent, but usually when the kernel (template) approaches 1/4-1/2 the size of the image frequency domain ...


3

Though the Matched Filter is the best tool detection of a known signals under AWGN it should work well here as well. To say something about the probabilities the question is, do you know something about the energy of the received signals? If you do, you should easily say something about the probabilities. Pay attention that if the assumption is a signal ...


3

This would be a cumbersome way to detect heart beats (or the QRS complex), if that is what you are trying to do ultimately. A little bit about what you are trying to do currently: Your observations are correct and to these I would like to add that no two heart beats are the same and therefore, strictly speaking, your template will be aligning just with ...


3

Here's the basic idea of what I know can be done, based on a talk by Professor Anurag Mittal of IIT Madras. The idea is of shape based object detection, but can obviously be extended elsewhere as well. Compute edgels using Berkeley edge detector. Connect edges obtained. "Global Object Boundary Detection". Shape matching using Chamfer distance or Houstoff ...


3

Yes, this should be enough for a basic isolated word recognition system. Probably not something for a commercial product, but good enough for a university project or demo... It would be better to ask the user to record a word and match against this, rather than attempt to match against a large database of utterances of the same word by different speakers. ...


2

Normalized is preferred to regular correlation to handle variation in brightness, or intensity, between images if nothing else. Regarding your second comment, the correlation is going to be lower if the shape rotates (unless it happens to be symmetrical in all four directions). There are other features you can consider that are rotation invariant as well. ...


2

I personally like this one. It is nicely designed for occlusions and large displacements. But, recent trends in deep learning lead to better results in optical flow such as this and this.


2

Vladimir: You're trying to detect a certain sample pattern (a certain sequence of sample values) within a longer discrete sequence of input samples. This sounds liked a "matched filter" problem to me. If the pattern of samples you're trying to detect is the samples in the red shaded region of your diagram then performing convolution seems like a smart thing ...


2

What's the maximum residual frequency offset allowable when doing the despreading of the CA code? 500 Hz? Or, what's the SNR loss due as a function of frequency missmatch due to residual Doppler? The maximum frequency offset is dependent on SNR required for acquisition as the roll-off of correlation versus frequency offset is due entirely to the duration of ...


2

That's hard. I'm not even sure I can do this myself before I actually try. It is definitely not something that can be described in a DSP.SE answer. About canny edge even if it is precise you will just get an image like a line drawing as an output. The point is what's next. There are many things to try the are many theory about pattern matching from the edge ...


2

A simple count of the pixels where the binary color is matching between the image to be recognized and the images of the ten reference digits might be good enough. For best results, it is advisable to create the reference images using exactly the same method as that for the target image. If it turns out that binarization doesn't work repeatably enough, you ...


2

I have referred to Stage I.D of this tutorial. Hope this helps. http://www.robots.ox.ac.uk/~vgg/practicals/instance-recognition/index.html#stage-id-improving-sift-matching-using-a-geometric-transformation When the features have scale and orientation assigned (e.g. SIFT features have these properties), you can compute similarity transform between each ...


1

What about an algorithm in image domain? For every pixel, check into the direction you desire to know (up-down in your images above, I guess) if there is a pixel with "the same" value (i.e. a value inside an acceptance band, of course. Let's say +/- 10%) as the pixel under investigation. If so, go further into that direction. Count every pixel that you find....


1

To increase the speed, I think the dimensionality of the feature descriptors should be reduced. SIFT feature descriptor is of 128 dimensions. So you can try PCA-SIFT which reduces the descriptor number to 20. It uses Principal Component Analysis to the SIFT descriptors. You can also try using Speeded Up Robust Features (SURF). The feature descriptors ...


1

Assuming that the tablature is well defined; let's say it's a MIDI file complete with pitch bend commands. you can make a synthesizer where each note is actually a tracking comb filter tuned to the midi note (and offset by the pitch bend command) with notches at each integer harmonic (including the fundamental). if those tracking comb filters processes a ...


1

You already converted to binary - if that proves to work well, then you can do a shrinkage operation (I'm quite sure there should be such in OpenCV but not completely) until you get one point per segment. Measure coordinates of these points and do pairwise comparison to array of coordinates from a set of models of the digits of a 7 segment display. ( first ...


1

Okay, Feature Extract and Matching is the way to go for a beginner. Other methods such as relative pose estimation, triangulation, 3d matching etc. would be much more complex. Your object is metallic and thus shiny. Also, without correct lighting it is quite difficult to match features. But still, as much as I could see from that picture, you have plentiful ...


1

I have not given it much thought, but I'm pretty sure a robust solution can be had without much trouble using classic Fourier Descriptors (FD). I think your problem might be a very good candidate for that. Don't think you need to do edge detection b/c you have black line drawings. Just start raster scanning until you hit any pixels, then do the following: ...


1

3. The polar image has the black areas, which are parts that can't be mapped to the original image because it would be outside of its boundaries. However, these areas have nearly disappeared in the log polar image. Why is that? Look at the second (polar) and third (log-polar) images. What's different about them? In the 3rd image, stuff that was ...


1

Start with edge detection. Then use Hough transform to detect straight lines. Then figure out which lines are parallel and which ones should be perpendicular. Then estimate the projective transformation so that the lines that should be perpendicular actually are perpendicular.


Only top voted, non community-wiki answers of a minimum length are eligible