|
Acadêmico(a): Anderson Dorow |
Título: Mecanismo de Busca Semântica de Áudio |
|
Abstract: |
This paper describes the development of an audio search engine based on its content. The mechanism obtains/receive an audio snippet, apply the Fourier transform to extract features from the audio, generates audio fingerprints, submit them to the server to find audio fingerprint matches. At the end of this process, there is a list of matches, which statistical methods are to define the degree of similarity between the audio input and those in the base. The results show that the similarity techniques implemented are sufficient to obtain positive results, even in the presence of noise and, the size of the input has influence in the efficiency of the search mechanism.
|
|
|