Solution
The issue is most likely connected with how IMAQ OCR Train.vi is used. For example, you may be using other VIs (such as IMAQ Read Text 4.vi) to recognize the objects, get character positions and then cropping objects using boundary coordinates and giving one letter as an input to the IMAQ OCR Train.vi. However, when you do this, characters that have low resolution are not recognized by the IMAQ OCR Train.vi. Hence, the empty string error. You should give the entire image and the ROI as an input to IMAQ OCR Train.vi, it would then iterate through each recognized object and train it as the corresponding character. Below you can see an example code of how this can be done.
