Description

In this paper, we present the design, implementation, and deployment of a wearable computing platform capable of automatically extracting and analyzing social-audio signals. Unlike conventional research that concentrates on data which have been recorded under constrained conditions, our data were recorded in completely natural and unpredictable situations. We design the WAM to continuously collect audio signals in completely natural and unpredictable situations. The proposed SASE architecture can be divided into three stages. Stage 1: Block detection of sound and speech and classification of both environmental and speech sounds. Stage 2: Voiced and non voiced speech segmentation and sound level meter calculation. Stage 3: Individual speaker segmentation, clustering, unknown number of speaker prediction, and social signal calculation. Load data: Load data is the process of collecting the information from the already trained audio signal. In this module, there are several number of audio signal which is taken as the dataset is trained by collecting the data in it. The data occurred from the each audio signal is displayed in the table format and the collected data is store as the ‘.mat’ file format. Input voice signal: In this process, the input audio file is in the format of ‘.wav’ file format. The input signal is also known as the raw signal. This raw signal consist of the several noise and the other unwanted speech signals also. So that the preprocessing is undertaken for the further enhancement of the input signal. Input voice signal: In this process, the input audio file is in the format of ‘.wav’ file format. The input signal is also known as the raw signal. This raw signal consist of the several noise and the other unwanted speech signals also. So that the preprocessing is undertaken for the further enhancement of the input signal. Segmentation is the process of separating the noise signal from the speaker signal. The segmentation consist of segmenting the data of the speaker from the outside noise. The speaker is nothing but the person who is given as the input for the process. Speech segmentation is a subfield of general speech perception and an important sub problem of the technologically-focused field of speech recognition, and cannot be adequately solved in isolation. Feature extraction is the process of collecting the information/data from the input voice signal. The feature is nothing but the data which consist of the several numerical values. The numerical value will be either the decimal or integer. This collection of information is saved as the ‘.mat’ format.