A cube-based saliency detection method using integrated visual and spatial features
发布日期：2016-12-01 18:16:26 阅读次数：次 作者：
核心提示：来源出版物: SENSOR REVIEW
|作者: Liu, T (Liu, Tao); Fang, ZX (Fang, Zhixiang); Mao, QZ (Mao, Qingzhou); Li, QQ (Li, Qingquan); Zhang, X (Zhang, Xing)|
|来源出版物: SENSOR REVIEW 卷: 36 期: 2 页: 148-157 DOI: 10.1108/SR-07-2015-0110 出版年: 2016|
摘要: Purpose - The spatial feature is important for scene saliency detection. Scene-based visual saliency detection methods fail to incorporate 3D scene spatial aspects. This paper aims to propose a cube-based method to improve saliency detection through integrating visual and spatial features in 3D scenes.
Design/methodology/approach - In the presented approach, a multiscale cube pyramid is used to organize the 3D image scene and mesh model. Each 3D cube in this pyramid represents a space unit similar to a pixel in the image saliency model multiscale image pyramid. In each 3D cube color, intensity and orientation features are extracted from the image and a quantitative concave-convex descriptor is extracted from the 3D space. A Gaussian filter is then used on this pyramid of cubes with an extended center-surround difference introduced to compute the cube-based 3D scene saliency.
Findings - The precision-recall rate and receiver operating characteristic curve is used to evaluate the method and other state-of-art methods. The results show that the method used is better than traditional image-based methods, especially for 3D scenes.Originality/value - This paper presents a method that improves the image-based visual saliency model.
联系地址： 中国·武汉市珞瑜路129号 邮编: 430079 E-mail：firstname.lastname@example.org