Biologically Motivated Object Detection Based on Integration of Top-Down and Bottom-Up AttentionReportar como inadecuado




Biologically Motivated Object Detection Based on Integration of Top-Down and Bottom-Up Attention - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Integration of top-down and bottom-up attention is crucial to developing a visual perception system. Existing attention models concentrate on bottom-up attention control, but lack of effective top-down attention guidance. We propose a new model that formalizes link between these two components. The bottom-up attention uses local scene representation to combine local complexity algorithm and linear filtering method. The top-down attention information is defined as a new holistic scene representation which works in three ways. The first is to discriminate between close-up and open scene categories. The second and the third are to provide reliable priors for the presence or absence of object and the location of it. Compared with traditional attention guidance algorithms, our algorithm shows how object detection is facilitated by scene classification and basing directly on entire scene without segmentation stages. Two new stages of pre-attention and focus attention enhance detecting performance and are more suitable for vision information processing in high level. Experiment results prove its effectiveness.

KEYWORDS

top-down attention; bottom-up attention; object detection; visual saliency; reature extraction

Cite this paper







Autor: Mei Tian, Siwei Luo, Jiali Zhao, Yaping Huang

Fuente: http://www.scirp.org/



DESCARGAR PDF




Documentos relacionados