Skip to content

Performance difference between vertical and horizontal defects on wood textures #55

Description

@huseyindedeyilmaz

I am currently training the EfficientAD model on my own dataset, which consists of walnut wood surfaces. These wood types typically have vertical grain patterns. I’ve noticed that the model struggles to detect vertically oriented defects, while it can easily detect very small and thin horizontal defects.

What could be the reason for this difference, and how can I improve the model’s performance on vertical defects?
I would really appreciate your help.

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