ABSTRACT
Coughing is a typical symptom of COVID-19. To detect and localize coughing sounds remotely, a convolutional neural network (CNN) based deep learning model was developed in this work and integrated with a sound camera for the visualization of the cough sounds. The cough detection model is a binary classifier of which the input is a two second acoustic feature and the output is one of two inferences (Cough or Others). Data augmentation was performed on the collected audio files to alleviate class imbalance and reflect various background noises in practical environments. For effective featuring of the cough sound, conventional features such as spectrograms, mel-scaled spectrograms, and mel-frequency cepstral coefficients (MFCC) were reinforced by utilizing their velocity (V) and acceleration (A) maps in this work. VGGNet, GoogLeNet, and ResNet were simplified to binary classifiers, and were named V-net, G-net, and R-net, respectively. To find the best combination of features and networks, training was performed for a total of 39 cases and the performance was confirmed using the test F1 score. Finally, a test F1 score of 91.9% (test accuracy of 97.2%) was achieved from G-net with the MFCC-V-A feature (named Spectroflow), an acoustic feature effective for use in cough detection. The trained cough detection model was integrated with a sound camera (i.e., one that visualizes sound sources using a beamforming microphone array). In a pilot test, the cough detection camera detected coughing sounds with an F1 score of 90.0% (accuracy of 96.0%), and the cough location in the camera image was tracked in real time.
ABSTRACT
OBJECTIVE: Experts disagree about the causes and significance of the recent increases in the prevalence of autism spectrum disorders (ASDs). Limited data on population base rates contribute to this uncertainty. Using a population-based sample, the authors sought to estimate the prevalence and describe the clinical characteristics of ASDs in school-age children. METHOD: The target population was all 7- to 12-year-old children (N=55,266) in a South Korean community; the study used a high-probability group from special education schools and a disability registry and a low-probability, general-population sample from regular schools. To identify cases, the authors used the Autism Spectrum Screening Questionnaire for systematic, multi-informant screening. Parents of children who screened positive were offered comprehensive assessments using standardized diagnostic procedures. RESULTS: The prevalence of ASDs was estimated to be 2.64% (95% CI=1.91-3.37), with 1.89% (95% CI=1.43-2.36) in the general-population sample and 0.75% (95% CI=0.58-0.93) in the high-probability group. ASD characteristics differed between the two groups: the male-to-female ratios were 2.5:1 and 5.1:1 in the general population sample and high-probability group, respectively, and the ratios of autistic disorders to other ASD subtypes were 1:2.6 and 2.6:1, respectively; 12% in the general-population sample had superior IQs, compared with 7% in the high-probability group; and 16% in the general-population sample had intellectual disability, compared with 59% in the high-probability group. CONCLUSIONS: Two-thirds of ASD cases in the overall sample were in the mainstream school population, undiagnosed and untreated. These findings suggest that rigorous screening and comprehensive population coverage are necessary to produce more accurate ASD prevalence estimates and underscore the need for better detection, assessment, and services.
