Manuscript received March 9, 2026; revised May 11, 2026; accepted June 30, 2026; published November 14, 2026.

Abstract—Lung auscultation is fast and low-cost, but its interpretation is subjective and sensitive to recording variability, leading to unidentified cases and elevated fatality rates. Automated screening is therefore necessary. Accurate identification of lung sound types remains challenging due to subtle acoustic variations and overlapping characteristics. The methods developed from current research rely on features produced by the Time Frequency Representations (TFR), which impact the discriminability and robustness. This study proposes Short-Time Variational Mode Decomposition (STVMD), a combination of the Short-Time Fourier Transform (STFT) with Variational Mode Decomposition (VMD) in two Strategies (STGs). Strategy 1 (STG 1) applies VMD-2D to the spectrograms, targeting the enhancement of 2D texture structures. While Strategy 2 (STG 2) applies VMD-1D to the waveform before spectrogram generation to improve TFR concentration through the pre-separation of 1D components. Gray-Level Co-occurrence Matrix (GLCM) and Gray-Level Difference Matrix (GLDM) features are extracted to be classified using a Support Vector Machine (SVM). On a five-class lung sound dataset, the method achieved the highest accuracy at 97.98% for STG 1 with substantially lower feature extraction time, 0.16 min. These results show that STG 1 offers the best accuracy-efficiency trade-off, enhancing discriminative spectrogram textures for GLCM-GLDM at minimal computational cost.

Keywords—auscultation, lung sound classification, short-time Fourier transform, variational mode decomposition

Cite: Annisa Humairani, Achmad Rizal, Salsabila Aurellia, and Rizkyandito A. Saputra, "Classification of Five Lung Sound Types Using STVMD and GLCM-GLDM-based Feature Extractors," Journal of Image and Graphics, Vol. 14, No. 6, pp. 974-986, 2026.

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).