A combination of extra audio features, curated training data, and agent-based label correction raises CA-SDRi from 11.088 dB to 12.726 dB on DCASE 2025 Task 4.
This section outlines the com- position of the official dataset and the specific data curation and augmentation steps to improve the performance of the model
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
eess.AS 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Performance improvement of spatial semantic segmentation with enriched audio features and agent-based error correction for DCASE 2025 Challenge Task 4
A combination of extra audio features, curated training data, and agent-based label correction raises CA-SDRi from 11.088 dB to 12.726 dB on DCASE 2025 Task 4.