Patient Sound Mel Spectrogram Dataset for Disease Detection

This dataset contains mel spectrogram representations of patient sound recordings, prepared for machine learning–based disease detection and classification tasks. The original data consists of WAV audio files capturing patient sounds associated with different health conditions. These audio recordings were preprocessed and converted into mel spectrograms to enable efficient use with modern deep learning architectures.Mel spectrograms were chosen as they provide a perceptually meaningful time–frequency representation of audio signals, closely aligned with human auditory perception and widely used in medical audio analysis, bioacoustics, and speech-based health diagnostics.The dataset is intended to support research in medical sound analysis, disease classification, audio-based diagnosis, and deep learning model benchmarking.