New Publication in JMIR Medical Informatics
PhD student Arifa Zahir published a paper on voice-based Parkinson disease screening in JMIR Medical Informatics.
PhD student Arifa Zahir and colleagues published the paper “A Machine Learning Approach to Voice-Based Parkinson Disease Screening Using Multiview Spectrogram and Speech Recognition Features: Diagnostic Study” in JMIR Medical Informatics.
The study presents a machine learning framework for noninvasive Parkinson disease screening using multiple spectrogram representations and speech-recognition-based features.
The proposed model combines Mel spectrogram, short-time Fourier transform, and constant-Q transform representations with a recognition-aware feature derived from automatic speech recognition.
박사과정 아리파 자히르(Arifa Zahir) 학생과 공동연구진의 논문 A Machine Learning Approach to Voice-Based Parkinson Disease Screening Using Multiview Spectrogram and Speech Recognition Features: Diagnostic Study가 JMIR Medical Informatics에 게재되었습니다.
본 연구에서는 음성 기반 파킨슨병 선별을 위해 다양한 스펙트로그램 표현과 음성인식 기반 특징을 결합한 머신러닝 프레임워크를 제안하였습니다.
제안한 방법은 Mel spectrogram, Short-Time Fourier Transform(STFT), Constant-Q Transform(CQT) 표현을 자동 음성인식(ASR) 기반 특징과 함께 활용하여 비침습적인 파킨슨병 선별 성능을 향상시켰습니다.