I’m using the OCR dataset that contains 30000 images of handwriting words in one folder. The ground truth labels are text files for these words and the classes are 40 characters. My concern is how to upload my dataset deep cognition and how to create the CSV file?
Hi so basically for input you have images and for output you have text files?
So for each input you have only 1 class output or different class?
Can you please share your input and output example for better understating?
Thanks for your fast respond.
The model will be a sequence model. Regarding the input will be an image of one Arabic word and the label will be its corresponding text.
The overall number of classes will be 35 letters which! are the Alphabet characters.
For each input word image that includes X characters, it will give X classes for the output as a sequence.
Sorry but right now DLS supports only English language as a output label.
But my label are English language not Arabic one.
this is one sequence label example:
So can I use DLS now?