Skip to content

normalization #149

@XIDIANPQZ

Description

@XIDIANPQZ

Hello, excellent work! I have a question regarding data preprocessing that I find a bit confusing. When processing the Human3.6M dataset (as seen in the following code: link), the values are normalized to the range [-1, 1]. However, during the preprocessing for "infer-wild" (referenced here: link), the scale is determined by the smaller dimension of (height, width). This results in a range that can exceed 1.

I’m wondering if this discrepancy in scaling could have an impact on the results. Could you clarify this?

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