Skip to content

Extract Paco_classifier for Local Pixelwise Classification #38

@kyrieb-ekat

Description

@kyrieb-ekat

Goal: Run the Paco_classifier pixelwise classifier locally (outside Rodan) to extract stafflines from music document images, using existing .h5 models produced by the previous Rodan deployment.

Source repository: DDMAL/Paco_classifier (https://github.com/DDMAL/Paco_classifier)

Reference implementation: fast_calvo_classifier.py in the Rodan repo — the run_my_task method shows the exact call pattern to replicate.

This is because Ich reckons and I agree we're going to end up needing this for staffline extraction and pitch finding. Using Rodan will be too time consuming at this stage--I'd like to extract the fast-pixelwise classifier so we can run it locally to knock out that phase of our experimentation and training.

The ensuing issues here are the steps and necessary things that will need to be done to do this successfully.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No fields configured for Task.

    Projects

    Status

    In progress

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions