Familiarity Based Unified Visual Attention Model for Fast and Robust Object Recognition

Joo-Young Kim, Kwanho Kim, Seungjin Lee, Minsu Kim, Hoi-Jun Yoo

Pattern Recognition |

Even though visual attention models using bottom-up saliency can speed up object recognition by predicting object locations, in the presence of multiple salient objects, saliency alone cannot discern target objects from the clutter in a scene. Using a metric named familiarity, we propose a top-down method for guiding attention towards target objects, in addition to bottom-up saliency. To demonstrate the effectiveness of familiarity, the unified visual attention model (UVAM) which combines top-down familiarity and bottom-up saliency is applied to SIFT based object recognition. The UVAM is tested on 3600 artificially generated images containing COIL-100 objects with varying amounts of clutter, and on 126 images of real scenes. The recognition times are reduced by 2.7×and 2×, respectively, with no reduction in recognition accuracy, demonstrating the effectiveness and robustness of the familiarity based UVAM.