An improved simplified PCNN model for salient region detection
- PDF / 3,121,047 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 76 Downloads / 188 Views
ORIGINAL ARTICLE
An improved simplified PCNN model for salient region detection Monan Wang1
· Xiping Shang1
Accepted: 4 November 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract As PCNN is modulated by using the pulse-coupled synaptic mechanisms, it has a great potential for image processing in a complex real-world environment, especially in images. A new simplified pulse coupled neural network (SPCNN) is proposed. This new model uses the pixel intensity with the actual physical meanings as the input parameters instead of the abstract network parameters in the original SPCNN. In order to achieve this goal, we try to derive the general formulae of dynamic threshold and internal activity of the SPCNN according to the dynamic properties of neurons and then deduce the relationship between the pixel intensity and the abstract parameters. Then, the relationship is transformed into an objective optimization problem to obtain the appropriate abstract parameters. Finally, extensive experiments are conducted on seven widely used datasets to demonstrate the effectiveness of the proposed method and shown improvement on the salient region detection. Keywords Pulse coupled neural network (PCNN) · Actual physical meanings · Pixel intensity · Salient region detection
1 Introduction The purpose of salient target detection [22,23] is to identify the most distinctive target or region in the image, and then segment it from the background. Unlike semantic segmentation and similar segmentation tasks, salient object detection pays more attention to very few interesting and attractive objects. Such a useful property allows salient object detection to commonly serve as the first step to a variety of computer vision applications including image segmentation [18,19,24– 26], image understanding [20] and feature extraction [21]. There have also been new developments in salient object detection in recent years, including video salient object detection [53,54] and RGB-D salient object detection [55–59]. Typical architectures of salient object detection methods can be divided into five categories: single-stream network, multi-stream network, side-fusion network, bottom-up/topElectronic supplementary material The online version of this article (https://doi.org/10.1007/s00371-020-02020-2) contains supplementary material, which is available to authorized users.
B
Monan Wang [email protected] Xiping Shang [email protected]
1
School of Mechanical and Power Engineering, Harbin University of Science and Technology, Harbin 150080, Heilongjiang, China
down network, and branched network [49]. Single-stream network is a standard architecture consisting of a sequential cascade of convolution layers, pooling layers, and nonlinear activation operations [50–52]. Multi-stream network typically has multiple network streams, each of which is trained with an input at a particular resolution to explicitly learn multiscale saliency features. The outputs from different streams are then combined together for the final prediction
Data Loading...
