Skip to content

README.md

Latest commit

 

History

History
47 lines (34 loc) · 1.23 KB

README.md

File metadata and controls

47 lines (34 loc) · 1.23 KB
Raw

Command Templates

This folder contains ready-to-adapt command templates for the main Graphics-LPIPS-QualCompare metric workflow.

Files:

  • train_metric.txt: train a model
  • evaluate_metric.txt: evaluate a checkpoint
  • correlate_metric.txt: compute correlation summaries

These templates are examples for users who want to run the metric on their own rendered data. They are not the paper revalidation pipeline.

Related documentation:

Placeholders

Adapt these values before running the commands:

  • dataset CSV paths under dataset/
  • --src_root, which must point to a QualCompare render root
  • model name passed with --name during training and -m during evaluation
  • database, render method, view method, and number of views

Expected rendered structure:

<EXPERIMENT_ROOT>/
  Source/
    <N>VP/
      <REFERENCE_OBJECT>/
        views/
        patchs/
  Distorted/
    <N>VP/
      <DISTORTED_OBJECT>/
        views/

The folder name is currently patchs, not patches.

The .txt examples use Windows PowerShell line continuations. For cmd.exe, replace backticks with ^; for Bash, replace them with \.