arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2410.22271eess.AScs.AIeess.IVeess.SP
收藏

利用混响和视觉深度线索进行带距离估计的声音事件定位与检测

Leveraging Reverberation and Visual Depth Cues for Sound Event Localization and Detection with Distance Estimation

University of Surrey(萨里大学)

Davide Berghi, Philip J. B. Jackson

AI总结:

针对DCASE2024任务3的带距离估计视听声音事件定位检测问题,提出基于AV-Conformer的多版系统,结合混响、深度等特征并采用集成策略,在基准数据集上实现性能提升。

摘要:

本报告介绍了我们提交至DCASE2024挑战赛任务3:带声源距离估计的音频与视听声音事件定位与检测(赛道B)的系统。我们的主模型基于视听(AV)Conformer(共视变换器),分别处理通过ResNet50提取的视频嵌入,以及通过在SELD(声音事件定位与检测)上预训练的音频编码器提取的音频嵌入。该模型在STARSS23数据集的开发集上大幅优于视听基线模型,将其DOAE(方向角误差)减半,且F1值提升了3倍以上。我们的第二个系统对AV-Conformer的输出进行时序集成。随后我们为模型扩展了距离估计特征,例如从全向音频通道提取的直达与混响信号分量,以及从视频帧提取的深度图。新系统虽然将我们此前模型的RDE(相对距离误差)提升了约3个百分点,但F1得分更低。分析表明,这可能是由训练集中极少出现的声音类别导致的,更复杂的系统无法检测到这些类别。为解决这一问题,我们的第四个也是最终系统采用了一种集成策略,结合了另外三个系统的预测结果。未来的消融实验可以测试许多优化系统与训练策略的方向,有望在这一听觉任务上实现性能的逐步提升。

英文摘要:

This report describes our systems submitted for the DCASE2024 Task 3 challenge: Audio and Audiovisual Sound Event Localization and Detection with Source Distance Estimation (Track B). Our main model is based on the audio-visual (AV) Conformer, which processes video and audio embeddings extracted with ResNet50 and with an audio encoder pre-trained on SELD, respectively. This model outperformed the audio-visual baseline of the development set of the STARSS23 dataset by a wide margin, halving its DOAE and improving the F1 by more than 3x. Our second system performs a temporal ensemble from the outputs of the AV-Conformer. We then extended the model with features for distance estimation, such as direct and reverberant signal components extracted from the omnidirectional audio channel, and depth maps extracted from the video frames. While the new system improved the RDE of our previous model by about 3 percentage points, it achieved a lower F1 score. This may be caused by sound classes that rarely appear in the training set and that the more complex system does not detect, as analysis can determine. To overcome this problem, our fourth and final system consists of an ensemble strategy combining the predictions of the other three. Many opportunities to refine the system and training strategy can be tested in future ablation experiments, and likely achieve incremental performance gains for this audio-visual task.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑