MAGENTA: Magnitude and Geometry-ENhanced Training Approach for Robust Long-Tailed Sound Event Localization and Detection
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866909797539905536 |
|---|---|
| author | Yeow, Jun-Wei Tan, Ee-Leng Peksi, Santi Gan, Woon-Seng |
| author_facet | Yeow, Jun-Wei Tan, Ee-Leng Peksi, Santi Gan, Woon-Seng |
| contents | Deep learning-based Sound Event Localization and Detection (SELD) systems degrade significantly on real-world, long-tailed datasets. Standard regression losses bias learning toward frequent classes, causing rare events to be systematically under-recognized. To address this challenge, we introduce MAGENTA (Magnitude And Geometry-ENhanced Training Approach), a unified loss function that counteracts this bias within a physically interpretable vector space. MAGENTA geometrically decomposes the regression error into radial and angular components, enabling targeted, rarity-aware penalties and strengthened directional modeling. Empirically, MAGENTA substantially improves SELD performance on imbalanced real-world data, providing a principled foundation for a new class of geometry-aware SELD objectives. Code is available at: https://github.com/itsjunwei/MAGENTA_ICASSP |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_15599 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MAGENTA: Magnitude and Geometry-ENhanced Training Approach for Robust Long-Tailed Sound Event Localization and Detection Yeow, Jun-Wei Tan, Ee-Leng Peksi, Santi Gan, Woon-Seng Audio and Speech Processing Sound Deep learning-based Sound Event Localization and Detection (SELD) systems degrade significantly on real-world, long-tailed datasets. Standard regression losses bias learning toward frequent classes, causing rare events to be systematically under-recognized. To address this challenge, we introduce MAGENTA (Magnitude And Geometry-ENhanced Training Approach), a unified loss function that counteracts this bias within a physically interpretable vector space. MAGENTA geometrically decomposes the regression error into radial and angular components, enabling targeted, rarity-aware penalties and strengthened directional modeling. Empirically, MAGENTA substantially improves SELD performance on imbalanced real-world data, providing a principled foundation for a new class of geometry-aware SELD objectives. Code is available at: https://github.com/itsjunwei/MAGENTA_ICASSP |
| title | MAGENTA: Magnitude and Geometry-ENhanced Training Approach for Robust Long-Tailed Sound Event Localization and Detection |
| topic | Audio and Speech Processing Sound |
| url | https://arxiv.org/abs/2509.15599 |