This is pretty funny because modern ECG machines generate digital output. You might first look into whether you can get the digits directly from the machine rather than scans of paper.
But suppose you can't. I assume you asking how to find numerical coordinates for the curve by scanning and then analysing the scan. The process of converting a pixmap or bitmap--a big 2d array of pixel data--to polyline coordinates is called "vectorzation." I did a quick Google search and came up with this page, which mentions several open source vectorizers: http://wiki.inkscape.org/wiki/index.php/Tools Most of these tools are going to produce SVG files. You'll probably need to look up the SVG file format to determine how to extract the coordinates you need. Since you are training neural nets, you are going to have to "clean" the results: put them all on the same vertical and horizontal scale, ensure the grid in the background isn't interpreted as data, etc. If you are required to implement the vectorizer yourself, write back. I have done that and can give you pointers. Once you have numerical descriptions of the code, I'd stay in MATLAB. Just so you know, professional ECG analyzers usually work in the _frequency_ domain. I.e. they work on the DFT of the ECG signal, not the signal itself. Or they use both, but the DFT is primary. With MATLAB, taking the DFT is one line of code. And MATLAB has a neural net toolkit. On Jan 8, 1:08 am, Deepak Garg <deepakgarg...@gmail.com> wrote: > my question is about "how to achieve digitization of an image/graph image". > for example i have the following ECG image( taken from a normal camera ):- > > http://i.stack.imgur.com/QAMfk.png > > so what algorithm should i follow to get the digitized image, my final aim > is to feed this information to a neural network that can classify the given > ECG image into a disease class. please suggest me which platform is more > feasible MATLAB or JAVA? > > please help me guys!! -- You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To post to this group, send email to algogeeks@googlegroups.com. To unsubscribe from this group, send email to algogeeks+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/algogeeks?hl=en.
