Spatial relationship code not running on windows machine

[KKcorps]

I'm trying to execute the layout file in command line but it simply doesn't run on windows. I haven't tried on Linux. Are there any dependencies needed for this to execute the c++ code?

[rlaz]

Dear Kartik,

Thanks for your interest in our system.

The dependencies for the system should be described in the documentation. We ran the system on Linux machines - running on Windows unless you use Cygwin or something similar may be challenging.

This is a busy time, so I may not be very quick to respond, but do let me know if you have other questions.

Best wishes, Prof. Zanibbi

On 4/3/16 8:35 AM, Kartik Khare wrote:

I'm trying to execute the layout file in command line but it simply doesn't run on windows. I haven't tried on Linux. Are there any dependencies needed for this to execute the c++ code?

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/DPRL/CROHME_2014/issues/1

Richard Zanibbi Associate Professor, Department of Computer Science Director, Document and Pattern Recognition Lab (dprl) Rochester Institute of Technology, NY, USA http://www.cs.rit.edu/~rlaz

dprl: http://www.cs.rit.edu/~dprl

"Think wrongly, if you please, but in all cases think for yourself." (Doris Lessing)

[KKcorps]

Just compiled all the c++ files into a .exe and the spatial relationship svm worked perfectly. But most of the time is gets confused between subscript, superscript and horizontal but inside and below get recognised very well. Tried changing the normalizing parameter from height of first symbol to euclidean distance between the centers of symbols but results still remained the same.

[rlaz]

I would expect some confusions between subscript, superscript and horizontal adjacent symbols, but not 'mostly' confusions.

Which dataset are you running this on?

On 4/6/16 12:20 AM, Kartik Khare wrote:

Just compiled all the c++ files into a .exe and the spatial relationship svm worked perfectly. But most of the time is gets confused between subscript, superscript and horizontal but inside and below get recognised very well. Tried changing the normalizing parameter from height of first symbol to euclidean distance between the centers of symbols but results still remained the same.

— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/DPRL/CROHME_2014/issues/1#issuecomment-206112982

Richard Zanibbi Associate Professor, Department of Computer Science Director, Document and Pattern Recognition Lab (dprl) Rochester Institute of Technology, NY, USA http://www.cs.rit.edu/~rlaz

dprl: http://www.cs.rit.edu/~dprl

"Think wrongly, if you please, but in all cases think for yourself." (Doris Lessing)

[KKcorps]

I'm testing on CROHME 2012 trainData dataset. I wrote my own code for testing in which I take a .inkml file as input, then extract the symbol traces from it, group and normalize the symbols and pass the symbol list to the Minimum Spanning Tree function. Am I doing something wrong? Do I need to pre process the traces from the .inkml in some way before grouping them?

[rlaz]

I suggest using the provided code, which does pre-process (i.e. modify) the stroke data, and see if your results change after that.

On 4/8/16 11:36 AM, Kartik Khare wrote:

I'm testing on CROHME 2012 trainData dataset. I wrote my own code for testing in which I take a .inkml file as input, then extract the symbol traces from it, group and normalize the symbols and pass the symbol list to the Minimum Spanning Tree function. Am I doing something wrong? Do I need to pre process the traces from the .inkml in some way before grouping them?

— You are receiving this because you commented. Reply to this email directly or view it on GitHub https://github.com/DPRL/CROHME_2014/issues/1#issuecomment-207483016

Richard Zanibbi Associate Professor, Department of Computer Science Director, Document and Pattern Recognition Lab (dprl) Rochester Institute of Technology, NY, USA http://www.cs.rit.edu/~rlaz

dprl: http://www.cs.rit.edu/~dprl

"Think wrongly, if you please, but in all cases think for yourself." (Doris Lessing)