Experimental results reveal that human perception tends to ignore subtle or trivial events in the event semantic identification, while model inference is easily affected by events with noises. Meanwhile, in event existence detection, models are usually more sensitive than humans.
Sep 10, 2024
This paper constructs a Multi-Annotated Foreground Audio Event Recognition (MAFAR) dataset, which comprises audio recordings labelled by 10 professional ...
Sep 13, 2024 · Experimental results reveal that human perception tends to ignore subtle or trivial events in the event semantic identification, while model ...
Sep 10, 2024 · The paper examines the differences between how humans perceive and interpret audio events compared to how AI models infer and recognize those same audio events.
To address this issue, this paper introduces the concept of semantic importance in AER, focusing on exploring the differences between human perception and model ...
Audio Event Recognition (AER) traditionally focuses on detecting and identifying audio events. Most existing AER models tend to detect all potential events ...
为了解决这一问题，本文引入了AER中的语义重要性概念，重点探讨人类感知与模型推断之间的差异。本文构建了一个多标注前景音频事件识别（MAFAR）数据集，该数据集包含由10名 ...
By comparing human annotations with the predictions of ensemble pre-trained models, this paper uncovers a significant gap between human perception and model ...
Experimental results reveal that human perception tends to ignore subtle or trivial events in the event semantic identification, while model inference is easily ...
Exploring Differences between Human Perception and Model Inference in Audio Event Recognition · Yizhou TanYanru Wu +5 authors. M. Plumbley. Computer Science.