首页|Researcher at VIT-AP University Publishes New Study Findings on Machine Learning (Emergency Vehicle Classification Using Combined Temporal and Spectral Audio Fe atures with Machine Learning Algorithms)
Researcher at VIT-AP University Publishes New Study Findings on Machine Learning (Emergency Vehicle Classification Using Combined Temporal and Spectral Audio Fe atures with Machine Learning Algorithms)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News – Investigators discuss new findings in artificial intelligence. According to news reporting originating from Andhra Pradesh, India , by NewsRx correspondents, research stated, “This study presents a novel approa ch to emergency vehicle classification that leverages a comprehensive set of inf ormative audio features to distinguish between ambulance sirens, fire truck sire ns, and traffic noise. A unique contribution lies in combining time domain featu res, including root mean square (RMS) and zero-crossing rate, to capture the tem poral characteristics, like signal energy changes, with frequency domain feature s derived from short-time Fourier transform (STFT).” The news editors obtained a quote from the research from VIT-AP University: “The se include spectral centroid, spectral bandwidth, and spectral roll-off, providi ng insights into the sound’s frequency content for differentiating siren pattern s from traffic noise. Additionally, Mel-frequency cepstral coefficients (MFCCs) are incorporated to capture the human-like auditory perception of the spectral i nformation. This combination captures both temporal and spectral characteristics of the audio signals, enhancing the model’s ability to discriminate between eme rgency vehicles and traffic noise compared to using features from a single domai n. A significant contribution of this study is the integration of data augmentat ion techniques that replicate real-world conditions, including the Doppler effec t and noise environment considerations. This study further investigates the effe ctiveness of different machine learning algorithms applied to the extracted feat ures, performing a comparative analysis to determine the most effective classifi er for this task. This analysis reveals that the support vector machine (SVM) ac hieves the highest accuracy of 99.5 %, followed by random forest (RF ) and k-nearest neighbors (KNNs) at 98.5%, while AdaBoost lags at 9 6.0% and long short-term memory (LSTM) has an accuracy of 93% .”
NETL
NSTL
