|
Acadêmico(a): André Felipe de Almeida |
Título: Sistema para Classificação de Objetos Sonoros |
|
Abstract: |
To identify objects by sound requires some computational processing steps, as follows: acquisition of audio to be recorded, feature extraction, registration and identificate information collected from the data reported. This paper presents a software to perform the identification of objects using the sounds they produce. Obtaining audio is done through the use of microphone or audio files. The sound feature extraction is performed using Fourier transform. For identificating objects similarities algorithms are used. The results demonstrate that the approach chosen have not brought good results. Despite the similarities algorithms have shown good results in music recognition, comparison of features has low accuracy, considering the amount of extracted features.
|
|
|