I use ABINET to train text recognition (including English case, digits and special characters),without using the pre-trained model, The dictionary uses 'DICT91', but in the actual recognition effect, the upper case of English is identified as lower case. Please tell me how to solve this problem, thank you.
I use ABINET to train text recognition (including English case, digits and special characters),without using the pre-trained model, The dictionary uses 'DICT91', but in the actual recognition effect, the upper case of English is identified as lower case. Please tell me how to solve this problem, thank you.