Towards Robust Transcription

Supplementary Materials for ISMIR 2024 LBD

Y. Kim and A. Lerch, "Towards Robust Transcription: Exploring Noise Injection Strategies for Training Data Augmentation," accepted for presentation at the Late-Breaking Demo Session of the 25th International Society for Music Information Retrieval Conference (ISMIR), San Francisco, USA, 2024. [arXiv]

Code

• plot_results.ipynb: Generates plots for the figures in the paper.
• significance_test.ipynb: Performs t-tests to assess the significance of model performance differences between clean data (CNR=∞) and perturbed datasets (CNR={0, 1/3, 1, 3}).

Results

• inference_results.md: A Markdown file containing the inference results for Figures 1 and 2 from the paper.
• inference_results.tex: A LaTeX file presenting the inference results for Figures 1 and 2 from the paper.
• inference_results.pdf: A PDF-rendered version of inference_results.tex.
• significance_test_results.md: A Markdown file presenting t-test results comparing clean (CNR = ∞) and perturbed datasets (CNR = {0, 1/3, 1, 3}).
• significance_test_results.tex: A LaTeX file presenting t-test results comparing clean (CNR = ∞) and perturbed datasets (CNR = {0, 1/3, 1, 3}).
• significance_test_results.pdf: A PDF-rendered version of significance_test_results.tex.

Contact

• Yonghyun Kim
  Email: yonghyun.kim@gatech.edu