DCASE2016: Acoustic Scene and Event Detection Challenge Submissions

A collection of system outputs and technical reports submitted for the DCASE2016 Challenge, archived by Annamaria Mesaros of Tampere University. The package enables later evaluation with additional metrics and is documented in a 2018 IEEE/ACM journal article. It contains the results and descriptions of all participating systems.

Use Cases

• Benchmarking new audio classification metrics based on the archived system outputs.
• Analyzing historical trends in acoustic event detection system design based on the technical reports.
• Developing meta-learning approaches for audio tasks using the collection of diverse system results.
• Reproducing or extending the original DCASE2016 challenge evaluation.

Strengths

• Archives the complete outcome of a major international audio challenge (DCASE2016).
• Includes both system outputs and technical reports, providing a holistic view of the challenge.
• Has a formal, peer-reviewed publication (IEEE/ACM TASLP 2018) documenting the challenge and this package.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Tampere University (Annamaria Mesaros et al.) via paperswithcode.

Collection Method

Aggregated submissions to the DCASE2016 Challenge.