The content of this work is based on the characteristics of standard artificial bee colony(ABC) algorithm with weak local search ability and slow convergence speed. Then, an improved algorithm named KD-ABC is proposed. For improving the diversity and quality of the solution, it changes the generation method of honey source. In the initialization phase, it uses the cluster center generated by the K-MEANS method as the initial honey source instead of the initialization in the standard method. For improving the local optimization ability and the convergence speed without reducing the global search, we proposed a dynamic neighborhood search mechanism based on the number of iterations in terms of ABC search strategy and neighborhood selection stage. In order to find a suitable threshold to divide the grayscale image into blood vessels and background parts, we applied the characteristics of the KD-ABC algorithm to the binary processing stage of the fundus retinal blood vessel image, which lays the foundation for future image recognition.
Cognitive wireless sensor networks (CWSNs) can use the idle authorized frequency band to solve the problem of spectrum resource shortage in traditional wireless sensor network. By employing spectrum hole in the authorized frequency band, the spectrum sensing technology can degrade the coexistent interference and enhance the performance of whole sensor network. Due to the characteristics of limited battery energy and low processing capacity with sensor nodes, it is necessary to enhance the energy efficiency while improving spectrum sensing performance. In this paper, a cooperative spectrum sensing strategy for CWSNs based on particle swarm optimization is proposed. Firstly, the system throughput and energy consumption are quantitatively analyzed, and the mathematical model related to energy efficiency is established. Secondly, the particle swarm optimization (PSO) algorithm is used to obtain the optimal selected nodes set under the limited conditions of false alarm probability and detection probability. To avoid local optimization in the process of problem solving, Cauchy mutation method is introduced to optimize the parameter selection of fitness function. The experimental results illustrate that our proposed method can improve the throughput of the system while ensuring the sensing performance, and achieve the energy efficiency effectively.
Text detection is the premise and guarantee of text recognition. Multi-oriented text detection is the current research hotspot. Due to the variability in size, spatial layout, color and the arrangement direction of natural scene text, natural scene text detection is still very challenging. Therefore, this paper proposes a simple and fast multi-oriented text detection method. Our method first optimizes the regression branch by designing a diagonal adjustment factor to make the position regression more accurate, which increases F-score by 0.8. Secondly, we add an attention module to the model, which improves the accuracy of detecting small text regions and increases F-score by 1.2. Then, we introduce DR Loss to solve the problem of positive and negative sample imbalance, which increases F-score by 0.5. Finally, we conduct experimental verification and analysis on the ICDAR2015, MSRA-TD500 and ICDAR2013 datasets. The experimental results demonstrate that this method can significantly improve the precision and recall of scene text detection, and it has achieved competitive results compared with existing advanced methods. On the ICDAR 2015 dataset, the proposed method achieves an F-score of 0.849 at 9.9fps at 720p resolution. On the MSRA-TD500 dataset, the proposed method achieves an F-score of 0.772 at 720p resolution. On the ICDAR 2013 dataset, the proposed method achieves an F-score of 0.887 at 720p resolution.
The iris is an area of the approximation of the ring between the pupil and the sclera of the human eye, which contains a large number of texture features. Due to the stability and specificity of the iris texture features, the iris can solve problems such as racial classification. The existing method of racial classification by iris image mainly adopts manual extraction of features and classification research, which has certain limitations. We proposed a method based on improved residual network for iris image race classification. In order to more fully extract the iris image features, we divide the network into two parts. We take apart in the first part of the network to the channel, for each channel for unused convolution kernels is utilized to extract features, and then connect the second part the back-end network residual, need special pointed out that in order to increase the receptive field. We use dilate convolution in each convolution layer, and we also use CAIAS and UBIRIS public data sets to verify the effectiveness of our method, the classification accuracy is 96.71%, and F1-score is 0.97. It is a good way to realize the basic classification of species and subspecies, improves the precision and feasibility of racial classifications, makes the racial classification theory more perfect.INDEX TERMS Iris images, racial classifications, dilate convolution, residual network.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.