Abstract Extraction from Audios
Abstract
Abstract extraction from audios is the method of identifying the beneficial and essential information from a given audio. This process involves speech signals to be converted into a sequence of words, other linguistic units by making use of an algorithm which is implemented as a computer program. An extractive speech abstraction method is devised on the obtained sequence of words. Extractive speech abstraction normally works as a binary classifier determining whether a sentence is a part of the abstract or not. As a result, using rank based classifiers, we represent the importance among sentences. Based on the weights assigned, the sentences are ranked. Consequently, a high-quality abstract is extracted from the input by choosing the sentences which are highly ranked and formed extract is stored as a text.
Keywords:abstract, text, document, ranking, sentence