Visual Saliency Detection in Natural Scene Images using a Selective Attention Model
Obtención de rasgos preponderantes en imágenes de escena natural mediante un Modelo de Atención Selectiva
| 2017 | |
|
Obtención de rasgos preponderantes Atención selectiva Filtro de Gabor en 2D Agrupamiento de datos Red neuronal de competencia Visual saliency detection Selective attention 2D Gabor filter Data clustering Competitive neural network |
|
| Español | |
| Universidad de Montevideo | |
| REDUM | |
|
http://revistas.um.edu.uy/index.php/ingenieria/article/view/306
https://hdl.handle.net/20.500.12806/2481 |
|
| Acceso abierto | |
| Atribución 4.0 Internacional |
| Sumario: | This article presents visual saliency detection in natural scene images. Images are processed using RGB, HSI and CMY color models and use some combinations of color components to feed a selective attention model based on the application of a two-dimensional specialized Gabor filter, which gives some of the features (like edges and outstanding contrasts), to be later highlighted by a clustering stage and a competitive artificial neural network stage. The simulations results show that the system is able to perform visual saliency detection in simple scenes and show encouraging results in complex scenes. For the tests were used images in RGB color format of 640 × 480 pixels (VGA). The implementation was made in the MATLAB® language. |
|---|
