Identifying Depression in a Person Using Speech Signals by Extracting Energy and Statistical Features

In present scenario, increasing state of anxieties, mental issues have become a common problem. A person having this issue cannot focus on regular tasks and it affects the mental state and leads to many issues such as suicide, anxiety attacks, personality disorders, eating disorders, and trauma related disorders. But if a person gets to know about it then it can be solved or reduced. This paper proposes a method to identify depression in subject based on its speech. This is done using feed forward neural network whic is a deep supervised learning method used to differentiate between things and finding a pattern based on the data like speech, images, numbers etc. and is used in this paper to classify whether a person is depressed or not by extracting energy and statistical features out of speech signals. This research can be helpful in easy diagnosis of depression so that victim can find cure and it can prevent suicides, mental trauma and other severe disorders due to prolonged suffering.