Search

The recognition of wood species is a laborious process, which is performed by experts, who attempt to distinguish the different species in different wood sections based on their macroscopic and microscopic characteristics. Most of these characteristics can be observed in the transverse or cross section of woods. According to experts, the next most important surface for wood species recognition is the tangential section while significant information can also be obtained from the radial section of woods. Based on the recent advances in the area of computer vision and pattern recognition, most researchers have proposed imagebased approaches attempting to address the problem either in microscopic or macroscopic scale. The main limitation is that in many cases there are some features that are not visible in each wood section. Firstly, we examine the contribution of each section in wood species recognition using two different computer-based texture analysis methods. Furthermore, we compare wood species recognition methods for both grayscale images and colorscale images. Finally, we propose a novel fusion method and we demonstrate that wood species recognition accuracy can be increased by fusing features from different wood sections. For the evaluation of the proposed method, a dataset, namely "WOOD-AUTH", consisting of more than 4272 wood images of twelve common wood species, was used. [ABSTRACT FROM AUTHOR]

Salient object detection has achieved great improvements by using the Fully Convolutional Networks (FCNs). However, the FCN-based U-shape architecture may cause dilution problems in the high-level semantic information during the up-sample operations in the top-down pathway. Thus, it can weaken the ability of salient object localization and produce degraded boundaries. To this end, in order to overcome this limitation, we propose a novel pyramid self-attention module (PSAM) and the adoption of an independent feature-complementing strategy. In PSAM, self-attention layers are equipped after multi-scale pyramid features to capture richer high-level features and bring larger receptive fields to the model. In addition, a channel-wise attention module is also employed to reduce the redundant features of the FPN and provide refined results. Experimental analysis demonstrates that the proposed PSAM effectively contributes to the whole model so that it outperforms state-of-the-art results over five challenging datasets. Finally, quantitative results show that PSAM generates accurate predictions and integral salient maps, which can provide further help to other computer vision tasks, such as object detection and semantic segmentation. [ABSTRACT FROM AUTHOR]