Salient Object Detection Based on Multiscale Segmentation and Fuzzy Broad Learning

Saliency detection has been a hot topic in the field of computer vision. In this paper, we propose a novel approach that is based on multiscale segmentation and fuzzy broad learning. The core idea of our method is to segment the image into different scales, and then the extracted features are fed to the fuzzy broad learning system (FBLS) for training. More specifically, it first segments the image into superpixel blocks at different scales based on the simple linear iterative clustering algorithm. Then, it uses the local binary pattern algorithm to extract texture features and computes the average color information for each superpixel of these segmentation images. These extracted features are then fed to the FBLS to obtain multiscale saliency maps. After that, it fuses these saliency maps into an initial saliency map and uses the label propagation algorithm to further optimize it, obtaining the final saliency map. We have conducted experiments based on several benchmark datasets. The results show that our solution can outperform several existing algorithms. Particularly, our method is significantly faster than most of deep learning-based saliency detection algorithms, in terms of training and inferring time.

Multiscale Optimized Segmentation of Urban Green Cover in High Resolution Remote Sensing Image

The urban green cover in high-spatial resolution (HR) remote sensing images have obvious multiscale characteristics, it is thus not possible to properly segment all features using a single segmentation scale because over-segmentation or under-segmentation often occurs. In this study, an unsupervised cross-scale optimization method specifically for urban green cover segmentation is proposed. A global optimal segmentation is first selected from multiscale segmentation results by using an optimization indicator. The regions in the global optimal segmentation are then isolated into under- and fine-segmentation parts. The under-segmentation regions are further locally refined by using the same indicator as that in global optimization. Finally, the fine-segmentation part and the refined under-segmentation part are combined to obtain the final cross-scale optimized result. The green cover objects can be segmented at their specific optimal segmentation scales in the optimized segmentation result to reduce both under- and over-segmentation errors. Experimental results on two test HR datasets verify the effectiveness of the proposed method.