Recognising images using fcdot templates

  • Last Post 14 March 2014
sreejith posted this 13 March 2014

I created a document definition (fcdot) using the Flexicapture engine SDK.
To keep things simple, I defined only a single page which contains 3 Anchors and 2 data fields

When I load this definition and try detecting images the following happens

  1. If I use the SAME image as I used for creating the definition, all the field positions are correct in the recognised result
  2. If I use any other image (even those which have very small difference in height and width), the field locations are offset by some numbers. All the fields are offset by the same number. The offsets depends on which Anchors I define

SDK_support posted this 14 March 2014

Dear Sreejith,

On the document definition file that you have sent us in your last letter some inaccuracy in matching anchors taking place, that is why this definition cannot be match on the other documents. As we can see the left bottom corner is absent on the image that you used for creating a form, instead of the corner there is a black area in the bottom of the picture. The left top square that you have matched as anchor too have rather complex boundaries, so it is not a very suitable element for anchor. Please, try to set as third anchor some static text, for example the full line “PAYBACK REGISTRATION FORM”.

Hope this information will be helpful.