Panoramic Image Saliency Detection by Fusing Visual Frequency Feature and Viewing Behavior Pattern

The panoramic images are widely used in many applications. Saliency detection is an important task for panoramic image processing. Traditional saliency detection algorithms that are originally designed for conventional flat-2D images are not efficient for panoramic images due to their particular viewing way. Based on this consideration, we propose a novel saliency detection algorithm for panoramic images by fusing visual frequency feature and viewing behavior pattern. By extracting the spatial frequency information in viewport domain and computing the center-surround contrast of them for the whole panoramic image, the visual frequency feature for saliency detection is accurately obtained. Further more, the context of user’s viewing behavior is integrated with visual frequency feature to generate the final saliency map. The experimental results show that the proposed algorithm is superior to the state-of-the-art algorithms when Pearson Correlation Coefficient (CC) is used as the evaluation metric.