I have the x-y cordinates of some online handwriting samples from which I am computing some statistical parameters using which I want to make an HMM based recognizer.
As HMM toolkit or HTK is orginally made for Speech recognition, so I am not able to understand how to perform online word recognition using HTK.
Can someone please help me to understand how to give my parameters as input to this toolkit and what output it will give ?????
As you say, HTK was developed for speech recognition. The HTK toolkit is a collection of special purpose programs that all work together.
Here is a version of the manual that describes what each program was designed for, including expected inputs and outputs.
I will warn you though, you will have an uphill battle trying to use HTK for handwriting recognition. It simply wasn't written with that in mind. The opening paragraph of the linked manual says:
HMMs can be used to model any time series and the core of HTK is similarly general-purpose. However, HTK is primarily designed for building HMM-based speech processing tools, in particular recognisers. Thus, much of the infrastructure support in HTK is dedicated to this task.
