Babies cry because they are ill or in pain, and at a time because they are hungry or sleepy. It is very difficult for parents, especially the new ones, to know why their child is letting out the whimper. Parents worry and are tensed, as they cannot help their little one.
To help parents, scientists have developed a new app that uses artificial intelligence for identifying normal and abnormal cries. The method can detect if the cry is a result of hunger, illness or something else.
It is a known fact that each baby’s cry is unique, but the cries share some common features if the reason for wailing is the same. The app uses an algorithm based on automatic speech recognition to detect and recognise the features of infant cries.
The team of scientists used a compressed sensing technique (a way to process large data efficiently) to analyse and classify the signals. The technique reconstructs a signal with the help of sparse data and is very helpful when the cries are recorded in a noisy environment.
The important and distinguishing features of the algorithm used for distinguishing cry are
- Independent of individual crier: the algorithm can be used in almost all practical situations as a way to recognise and classify various cry features to understand why the baby is crying and how urgent the cries are.
- The ‘cry recognition language’ algorithm can distinguish between normal and abnormal cries, even in noisy surroundings.
- The algorithm uses automatic speech recognition for recognition and detection.
- Features like loudness, pitch and timbre are used along with wavelengths for analyzing and then comparing them with the databases available. The database of cries is created with the help of experienced and neonatal nurses and caregivers.
Experts say that the various cry sounds can have a lot of health-related information. The differences in the sound signals carry the information as the features differ. The method used for recognition would not help parents and caregivers but also assist in areas of medicine where decision making completely depends on experience.
The app or the technology was developed to help deaf parents, to figure out what their child needs, as they cannot hear the different cry sounds of their babies. This technology could even help normal parents and so it is being used as a universal tool. It could also help parents who have a busy schedule and cannot have their watchful eyes on their kid 24/7.
This app can be even used for identifying autism in children at an earlier stage. Earlier detection of this condition could help in handling the situation better and the outcome could be better than diagnosing it in a later stage. Experts in the field suggest that unusual vocal patterns in infants can be an early cue for autism. The app is still in the initial stage. Scientists are still working on it to make it useful for detecting various physical and mental disorders in infants because early detection can help in treating them or reducing the after effects.