Bone Cement Removal Audio Monitoring Dataset

本数据集由莱比锡大学汉诺威信息处理研究所和捷克科学院地球物理研究所共同创建，包含与骨水泥侵蚀过程相关的音频信号和侵蚀深度数据。数据集由1150个样本组成，每个样本包含一个音频文件和对应的侵蚀深度剖面。音频文件采用38.4 kHz采样率录制，并通过Mel频率谱图进行预处理。数据集旨在为优化流体喷射参数提供支持，并通过状态空间模型实现预测性控制，以实现更精确的骨水泥侵蚀过程。

This dataset was jointly created by the Institute of Information Processing, Hannover of Leipzig University and the Institute of Geophysics, Czech Academy of Sciences. It contains audio signals and erosion depth data related to the bone cement erosion process. The dataset consists of 1150 samples, each including one audio file and its corresponding erosion depth profile. Audio files were recorded at a sampling rate of 38.4 kHz and preprocessed using Mel-frequency spectrograms. This dataset is intended to support the optimization of fluid jet parameters and enable predictive control via state-space models, thereby facilitating more precise control over the bone cement erosion process.

搜集汇总

数据集介绍

构建方式

Bone Cement Removal Audio Monitoring Dataset is meticulously crafted to simulate clinical scenarios where bone cement is removed from bone samples. Utilizing a pulsating fluid jet device, the dataset is generated by varying process parameters such as standoff distance, nozzle diameter, sonotrode frequency, fluid pressure, and robot arm velocity. The audio signals are captured using a high-resolution microphone, and the erosion profiles are measured using an optical microscope, ensuring a comprehensive dataset that correlates audio signals with the material erosion process.

特点

This dataset is unique in its approach to bone cement removal, employing a pulsating fluid jet device that offers a minimally invasive and cold technique. The dataset includes audio recordings and erosion profiles, providing a valuable resource for researchers and engineers. The audio signals are preprocessed as Mel Spectrograms, facilitating frequency distribution analysis. The dataset is split into training and testing subsets, with the S4D-Bio model achieving a high validation accuracy of 98.93% in predicting erosion profiles from audio data.

使用方法

To utilize the Bone Cement Removal Audio Monitoring Dataset, researchers can download it from the IEEE Dataport. The dataset includes ablation profiles and audio files recorded by a high-resolution microphone. The audio files are preprocessed as Mel Spectrograms for analysis. The dataset is split into training and testing subsets, with the S4D-Bio model achieving a high validation accuracy of 98.93% in predicting erosion profiles from audio data. Researchers can use this dataset to train and validate machine learning models, specifically State Space Models, for predicting erosion profiles based on audio signals. The dataset can also be used to study the effects of varying process parameters on bone cement removal and to develop new techniques for minimally invasive bone cement removal.

背景与挑战

背景概述

在骨水泥移除的领域中，传统方法往往耗时且不精确，对周围组织可能造成损伤。为了解决这一问题，Bone Cement Removal Audio Monitoring Dataset数据集的创建为研究提供了一种新型精确、微创且低温的骨水泥移除技术。该数据集由Melanie Schaller等人于2023年创建，旨在通过使用脉冲流体喷射装置，结合先进的音频监测技术，实现骨水泥的精确移除。该数据集在骨水泥移除领域具有重要意义，为临床手术提供了新的技术支持和数据基础。

当前挑战

Bone Cement Removal Audio Monitoring Dataset数据集面临的挑战主要包括：1) 所解决的领域问题的挑战：如何在微创手术中精确地移除骨水泥，同时减少对周围组织的损伤；2) 构建过程中所遇到的挑战：如何在喷射过程中实现实时监测，以及如何优化流体喷射参数。为了应对这些挑战，研究人员采用了音频信号监测技术，并通过状态空间模型（SSM）S4D-Bio对流体喷射参数进行动态优化。此外，为了提高模型在噪声环境下的鲁棒性，研究人员采用了噪声处理策略，如dropout和批量归一化。

常用场景

经典使用场景

在骨科手术中，骨水泥的去除是一个重要的步骤，传统的手动方法往往耗时、不够精确，且可能对周围组织造成损害。本研究提出了一种新型脉冲流体喷射技术，结合先进的音频监控技术，实现了精确、微创且低温的骨水泥去除。该技术通过音频信号监控和状态空间模型（SSM）S4D-Bio动态优化流体喷射参数，克服了溅射造成的可见性障碍。研究生成了一组全面的数据集，将各种过程参数和相应的音频信号与材料侵蚀相关联，使用SSM实现了对预测侵蚀过程的精确控制，准确率达到98.93%。该研究表明，脉冲流体喷射装置与先进的音频监控技术相结合，是精确去除骨水泥的高效工具。

实际应用

该数据集的实际应用场景包括骨科手术中骨水泥的去除。通过音频信号监控和状态空间模型，可以实现实时调整流体喷射参数，提高手术的精确性和安全性。此外，该技术还可以应用于其他需要微创、精确和低温操作的手术场景，如眼科学或心血管手术。

衍生相关工作

该数据集衍生了关于脉冲流体喷射技术和音频监控技术在生物医学手术中的应用研究。此外，该研究还探索了状态空间模型在动态系统建模中的应用，以及噪声处理策略在机器学习模型中的应用。这些研究成果为未来的生物医学工程研究提供了新的思路和方法。