Publications by authors named "Melaku N Getahun"
Front Artif Intell
February 2024
Article Synopsis
- Crying in infants serves as a crucial signal indicating various states like discomfort, hunger, or sickness, but caregivers often struggle to interpret these cues effectively.
- This study explores advanced audio feature representations, such as time-domain features (zero-crossing rate and root mean square), frequency-domain features (Mel-spectrogram), and time-frequency-domain features (Mel-frequency cepstral coefficients), to analyze infant cries for better understanding.
- The research employs machine learning classifiers, notably a random forest classifier, achieving an accuracy of 96.39% in identifying the meaning behind cries, surpassing current state-of-the-art methods.
