Thanks to visit codestin.com
Credit goes to github.com

Skip to content

The issue of poor performance in the deblurring task with a custom dataset. #40

@ziFan99

Description

@ziFan99

I would like to ask about my current task, which involves recognizing blurred shipping labels. The dataset I am using is self-generated using a blur generation algorithm on clear shipping labels. However, after training with your model, the results have not been satisfactory. I would like to know if your model is suitable for such a scenario (i.e., document blur, where details such as text and numbers on the shipping labels are missing), or could it be that my dataset is too small? Could you also share what the dataset size was in your deblurring tasks? I look forward to your reply.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions