# Frequency Dynamic Convolution: Frequency-Adaptive Pattern Recognition for Sound Event Detection

arXiv: https://arxiv.org/abs/2203.15296 · HTML: https://mlsanity.com/paper/2203.15296/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Mar 2022.

## Built on by (most cited first, 23 of 23)

- [Pushing the Frontier of Audiovisual Perception with Large-Scale Multimodal Correspondence Learning](https://mlsanity.com/paper/2512.19687/) (2025)
- [TACOS: Temporally-aligned Audio CaptiOnS for Language-Audio Pretraining](https://mlsanity.com/paper/2505.07609/) (2025)
- [AND: Audio Network Dissection for Interpreting Deep Acoustic Models](https://mlsanity.com/paper/2406.16990/) (2024)
- [FlexSED: Towards Open-Vocabulary Sound Event Detection](https://mlsanity.com/paper/2509.18606/) (2025)
- [Dynamic Convolutional Neural Networks as Efficient Pre-trained Audio Models](https://mlsanity.com/paper/2310.15648/) (2023)
- [CASTELLA: Long Audio Dataset with Captions and Temporal Boundaries](https://mlsanity.com/paper/2511.15131/) (2025)
- [VIFS: An End-to-End Variational Inference for Foley Sound Synthesis](https://mlsanity.com/paper/2306.05004/) (2023)
- [Dual Knowledge Distillation for Efficient Sound Event Detection](https://mlsanity.com/paper/2402.02781/) (2024)
- [Few-shot Long-Tailed Bird Audio Recognition](https://mlsanity.com/paper/2206.11260/) (2022)
- [MFF-EINV2: Multi-scale Feature Fusion across Spectral-Spatial-Temporal Domains for Sound Event Localization and Detection](https://mlsanity.com/paper/2406.08771/) (2024)
- [Auditory Intelligence: Understanding the World Through Sound](https://mlsanity.com/paper/2508.07829/) (2025)
- [SynSonic: Augmenting Sound Event Detection through Text-to-Audio Diffusion ControlNet and Effective Sample Filtering](https://mlsanity.com/paper/2509.18603/) (2025)
- [Sounding Highlights: Dual-Pathway Audio Encoders for Audio-Visual Video Highlight Detection](https://mlsanity.com/paper/2602.03891/) (2026)
- [Leveraging Language Model Capabilities for Sound Event Detection](https://mlsanity.com/paper/2308.11530/) (2023)
- [Infant Cry Detection Using Causal Temporal Representation](https://mlsanity.com/paper/2503.06247/) (2025)
- [NukesFormers: Unpaired Hyperspectral Image Generation with Non-Uniform Domain Alignment](https://mlsanity.com/paper/2503.07004/) (2025)
- [Revisiting SSL for sound event detection: complementary fusion and adaptive post-processing](https://mlsanity.com/paper/2505.11889/) (2025)
- [Cross-talk based multi-task learning for fault classification of machine system influenced by multiple variables](https://mlsanity.com/paper/2602.05146/) (2026)
- [Sub-Band Spectral Matching with Localized Score Aggregation for Robust Anomalous Sound Detection](https://mlsanity.com/paper/2603.13749/) (2026)
- [Towards Open World Sound Event Detection](https://mlsanity.com/paper/2605.03934/) (2026)
- [Domain-incremental audio classification using domain-specific experts and prototype classifier](https://mlsanity.com/paper/2606.22952/) (2026)
- [Semi-Supervised Sound Event Detection with Conditional Mixup and Embedding-Level Contrastive Loss](https://mlsanity.com/paper/2606.29901/) (2026)
- [Seeing Through WiFi: Lightweight Human Pose Estimation with Dynamic Kernel Attention](https://mlsanity.com/paper/2607.03196/) (2026)

## Abstract (arXiv)

2D convolution is widely used in sound event detection (SED) to recognize two dimensional time-frequency patterns of sound events. However, 2D convolution enforces translation equivariance on sound events along both time and frequency axis while frequency is not shift-invariant dimension. In order to improve physical consistency of 2D convolution on SED, we propose frequency dynamic convolution which applies kernel that adapts to frequency components of input. Frequency dynamic convolution outperforms the baseline by 6.3% in DESED validation dataset in terms of polyphonic sound detection score (PSDS). It also significantly outperforms other pre-existing content-adaptive methods on SED. In addition, by comparing class-wise F1 scores of baseline and frequency dynamic convolution, we showed that frequency dynamic convolution is especially more effective for detection of non-stationary sound events with intricate time-frequency patterns. From this result, we verified that frequency dynamic convolution is superior in recognizing frequency-dependent patterns.