ABSTRACT
Voice is an essential component of human communication, serving as a fundamental medium for expressing thoughts, emotions, and ideas. Disruptions in vocal fold vibratory patterns can lead to voice disorders, which can have a profound impact on interpersonal interactions. Early detection of voice disorders is crucial for improving voice health and quality of life. This research proposes a novel methodology called VDDMFS [voice disorder detection using MFCC (Mel-frequency cepstral coefficients), fundamental frequency and spectral centroid] which combines an artificial neural network (ANN) trained on acoustic attributes and a long short-term memory (LSTM) model trained on MFCC attributes. Subsequently, the probabilities generated by both the ANN and LSTM models are stacked and used as input for XGBoost, which detects whether a voice is disordered or not, resulting in more accurate voice disorder detection. This approach achieved promising results, with an accuracy of 95.67%, sensitivity of 95.36%, specificity of 96.49% and f1 score of 96.9%, outperforming existing techniques.
Subject(s)
Voice Disorders , Voice , Humans , Quality of Life , Voice Quality , Speech Acoustics , Voice Disorders/diagnosis , AcousticsABSTRACT
INTRODUCTION: Carcinoma cervix is a common gynecologic malignancy in India and is treated with radical chemoradiation where intracavitary brachytherapy (ICR) is an integral part. In ICR of cervix, the two-dimensional (2D) point-based dosimetry cervix is the most common method used in high-volume centers with rectal dose calculation at modified ICRU rectal point with rectal wire placement. The rectal dose measurement using this method underestimates the dose to the rectum, and rectal dose also varies with the type of applicator used. The aim of our study is to compare the rectal dose calculated by ICRU 38 method versus rectal dose calculated by the rectal wire method using Henschke applicator. MATERIALS AND METHODS: This is a single-institute, dosimetric comparison study done prospectively. Fifty patients were planned for ICR after 2D orthogonal radiograph-based, computer planning by iridium 192 high-dose rate remote afterloading technique after placing the appropriate Henschke applicator. The vaginal packing was done using sterile gauze with contrast material for defining the ICRU 38 rectal point, and a rectal wire was placed for the modified ICRU rectal point. Rectal doses were calculated by both the methods and compared. RESULTS: The modified ICRU rectal point recorded a lower rectal dose (mean of 25%) compared to ICRU 38 rectal point in the study patients. There were ten patients (20%) with either too much or too little contrast material which made the visualization of the rectal point and radiation planning difficult. P value by paired t-test method was 0.0001, which was statistically significant. CONCLUSION: The modified ICRU rectal point is easier to visualize than ICRU 38 method (100% vs. 80%) for dosimetry, but it underestimates the rectal doses when compared to ICRU 38 rectal point. There needs to be a correction factor applied (25% in our study for Henschke applicator) when evaluating the rectal doses calculated by rectal wire method, to reduce the rectal toxicity.