Abstract In order to efficiently utilize high-level information and depth in RGB-D saliency object detection, multi-level fusion is studied. Different from existing methods which ignore feature dilution the process of downward transmission, a interactive method designed compared with five advanced models through four evaluation indexes. The experimental results show that model this paper advanc...

To overcome the drawback of using only low-level features for the description of image content and to fill the gap between the perceptual property and semantic meaning, this chapter presents an object-based scheme and some object level techniques for image retrieval. According to a multi-layer description model, images are analyzed in different levels for progressive understanding, and this pro...

The cosmological mass function problem is analyzed in full detail in the case of 1D gravity, with analytical, semi-analytical and numerical techniques. The extended Press & Schechter [1] theory is improved by detailing the relation between smoothing radius and mass of the objects. This is done by introducing in the formalism the concept of a growth curve for the objects. The predictions of the ...

Slender flexible objects are ubiquitous in real-world circumstances. The existing object detection and segmentation algorithms have achieved high accuracy speed rigid objects, but the effect of slender is not ideal. Extreme aspect ratios dynamically changeable geometric appearance characterize to extent that it difficult locate them instance segmentation. In this paper, a new method based on co...