scholarly journals Multiscale Convolutional Neural Networks with Attention for Plant Species Recognition

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Xianfeng Wang ◽  
Chuanlei Zhang ◽  
Shanwen Zhang
Keyword(s):  
Neural Network ◽  
Plant Species ◽  
High Frequency ◽  
Species Recognition ◽  
Low Frequency ◽  
Plant Leaf ◽  
Network Training ◽  
Feature Based ◽  
The Rich ◽  
Frequency Features

Plant species recognition is a critical step in protecting plant diversity. Leaf-based plant species recognition research is important and challenging due to the large within-class difference and between-class similarity of leaves and the rich inconsistent leaves with different sizes, colors, shapes, textures, and venations. Most existing plant leaf recognition methods typically normalize all leaf images to the same size and then recognize them at one scale, which results in unsatisfactory performances. A novel multiscale convolutional neural network with attention (AMSCNN) model is constructed for plant species recognition. In AMSCNN, multiscale convolution is used to learn the low-frequency and high-frequency features of the input images, and an attention mechanism is utilized to capture rich contextual relationships for better feature extraction and improving network training. Extensive experiments on the plant leaf dataset demonstrate the remarkable performance of AMSCNN compared with the hand-crafted feature-based methods and deep-neural network-based methods. The maximum accuracy attained along with AMSCNN is 95.28%.

Machine Learning and Convolution Neural Network Approaches to Plant Leaf Recognition

2021 ◽  
pp. 234-248
Author(s):  
Rajesh K. V. N. ◽  
Lalitha Bhaskari D.
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Plant Species ◽  
Intelligent Systems ◽  
Human Life ◽  
Species Recognition ◽  
Convolution Neural Network ◽  
Plant Leaf ◽  
Network Approaches ◽  
Number Of Plant Species

Plants are very important for the existence of human life. The total number of plant species is nearing 400 thousand as of date. With such a huge number of plant species, there is a need for intelligent systems for plant species recognition. The leaf is one of the most important and prominent parts of a plant and is available throughout the year. Leaf plays a major role in the identification of plants. Plant leaf recognition (PLR) is the process of automatically recognizing the plant species based on the image of the plant leaf. Many researchers have worked in this area of PLR using image processing, feature extraction, machine learning, and convolution neural network techniques. As a part of this chapter, the authors review several such latest methods of PLR and present the work done by various authors in the past five years in this area. The authors propose a generalized architecture for PLR based on this study and describe the major steps in PLR in detail. The authors then present a brief summary of the work that they are doing in this area of PLR for Ayurvedic plants.

scholarly journals An Unconstrained Face Recognition Method Based on Siamese Networks

2021 ◽  
Author(s):  
Song CunLi ◽  
Shouyong Ji
Keyword(s):  
Neural Network ◽  
Face Recognition ◽  
Network Model ◽  
High Frequency ◽  
Recognition Accuracy ◽  
Recognition Rate ◽  
Low Frequency ◽  
Image Features ◽  
Frequency Features ◽  
Frequency Feature

Abstract It is aimed at the low accuracy and low efficiency of face recognition under unlimited conditions.In this paper, a Siamese neural Network model SN-LF (Siamese Network based on LBP and Frequency Feature perception) is designed based on the Local Binary Pattern (LBP) and the Frequency sensing model.Based on Siamese neural networks, the network adopts circular LBP algorithm and frequency feature perception to realize face recognition under unrestricted conditions.The LBP algorithm can eliminate the influence of light on the image and provide directional input to the network model at the same time.Frequency feature sensing divides the image features into low frequency features and high frequency features. The low frequency features are compressed in the Siamese neural network to increase the recognition efficiency of the network. At the same time, information is exchanged with the high frequency features, so that the target noise data can be eliminated while the feature data is retained.In this way, the recognition rate of the network is maintained, and the computing speed of the network is improved.Simulation experiments are carried out on standard face dataset CASIA-Webface and Yale-B, and compared with other network models. The experimental results show that the proposed SN-LF network structure can improve the recognition accuracy of the algorithm, and achieve a good recognition accuracy.

scholarly journals Compression of Phase-Only Holograms with JPEG Standard and Deep Learning

2018 ◽  
Vol 8 (8) ◽  
pp. 1258 ◽  
Author(s):  
Shuming Jiao ◽  
Zhi Jin ◽  
Chenliang Chang ◽  
Changyuan Zhou ◽  
Wenbin Zou ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
High Frequency ◽  
Critical Issue ◽  
Compression Process ◽  
Compression Scheme ◽  
Quality Degradation ◽  
Digital Format ◽  
Enormous Amount ◽  
Frequency Features

It is a critical issue to reduce the enormous amount of data in the processing, storage and transmission of a hologram in digital format. In photograph compression, the JPEG standard is commonly supported by almost every system and device. It will be favorable if JPEG standard is applicable to hologram compression, with advantages of universal compatibility. However, the reconstructed image from a JPEG compressed hologram suffers from severe quality degradation since some high frequency features in the hologram will be lost during the compression process. In this work, we employ a deep convolutional neural network to reduce the artifacts in a JPEG compressed hologram. Simulation and experimental results reveal that our proposed “JPEG + deep learning” hologram compression scheme can achieve satisfactory reconstruction results for a computer-generated phase-only hologram after compression.

scholarly journals Statistical regularities cause attentional suppression with target-matching distractors

2020 ◽  
Author(s):  
Dirk Kerzel ◽  
Stanislas Huynh Cong
Keyword(s):  
High Frequency ◽  
Target Location ◽  
Low Frequency ◽  
Target Display ◽  
Distractor Location ◽  
Target Feature ◽  
Statistical Regularities ◽  
Feature Based ◽  
The Impact ◽  
Target Matching

AbstractVisual search may be disrupted by the presentation of salient, but irrelevant stimuli. To reduce the impact of salient distractors, attention may suppress their processing below baseline level. While there are many studies on the attentional suppression of distractors with features distinct from the target (e.g., a color distractor with a shape target), there is little and inconsistent evidence for attentional suppression with distractors sharing the target feature. In this study, distractor and target were temporally separated in a cue–target paradigm, where the cue was shown briefly before the target display. With target-matching cues, RTs were shorter when the cue appeared at the target location (valid cues) compared with when it appeared at a nontarget location (invalid cues). To induce attentional suppression, we presented the cue more frequently at one out of four possible target positions. We found that invalid cues appearing at the high-frequency cue position produced less interference than invalid cues appearing at a low-frequency cue position. Crucially, target processing was also impaired at the high-frequency cue position, providing strong evidence for attentional suppression of the cued location. Overall, attentional suppression of the frequent distractor location could be established through feature-based attention, suggesting that feature-based attention may guide attentional suppression just as it guides attentional enhancement.

Multisource Image Fusion Based on DWT and Simplified Pulse Coupled Neural Network

2013 ◽  
Vol 457-458 ◽  
pp. 736-740 ◽  
Author(s):  
Nian Yi Wang ◽  
Wei Lan Wang ◽  
Xiao Ran Guo
Keyword(s):  
Neural Network ◽  
Image Fusion ◽  
High Frequency ◽  
Low Frequency ◽  
Discrete Wavelet ◽  
Cortical Model ◽  
Fusion Algorithm ◽  
Global Coupling ◽  
Maximum Selection ◽  
Pulse Synchronization

In this paper, a new image fusion algorithm based on discrete wavelet transform (DWT) and spiking cortical model (SCM) is proposed. The multiscale decomposition and multi-resolution representation characteristics of DWT are associated with global coupling and pulse synchronization features of SCM. Two different fusion rules are used to fuse the low and high frequency sub-bands respectively. Maximum selection rule (MSR) is used to fuse low frequency coefficients. As to high frequency subband coefficients, spatial frequency (SF) is calculated and then imputed into SCM to motivate neural network. Experimental results demonstrate the effectiveness of the proposed fusion method.

scholarly journals Forecasting Soil Moisture In The Soil Under The Caragana Shrubland Using Wavelet Analysis And NARX Neural Network

2021 ◽  
Author(s):  
Dong-Mei Bai ◽  
Zhong-Sheng Guo ◽  
Man-Cai Guo
Keyword(s):  
Neural Network ◽  
Soil Moisture ◽  
Wavelet Analysis ◽  
Soil Water ◽  
High Frequency ◽  
Sustainable Use ◽  
Low Frequency ◽  
Frequency Component ◽  
High Frequency Component ◽  
Narx Neural Network

Abstract Purpose: It is important for sustainable use of soil water resources to forecast soil moisture in forestland of water-limited regions. There are some soil moisture models. However, there is not a better method to forecast soil moisture.Methods: The change of soil moisture with time were investigated and the data of soil moisture were divided into a low frequency and a high frequency component using wavelet analysis, and then NARX neural network was used to build model I and model II. For model I, low frequency component was the input variable, and for model II, low frequency component and high frequency component were predicted.Results: the average relative error for model I is 3.5% and for model II is 0.3%. The average relative error of predicted soil moisture in100cm layer using model II is 0.8%, then soil water content in 40 cm and 200 cm soil depth is selected and the forecast errors are 4.9 % and 0.4 %.Using model II to predict soil water is well．Conclusion: Predicting soil water will be important for sustainable use of soil water resource and controlling soil degradation, vegetation decline and crop failure in water limited regions.

scholarly journals Medical Image Fusion Based on Sparse Representation and PCNN in NSCT Domain

2018 ◽  
Vol 2018 ◽  
pp. 1-12 ◽  
Author(s):  
Jingming Xia ◽  
Yiming Chen ◽  
Aiyue Chen ◽  
Yicai Chen
Keyword(s):  
Neural Network ◽  
Image Fusion ◽  
Sparse Representation ◽  
High Frequency ◽  
Medical Image ◽  
Information Transfer ◽  
Color Image ◽  
Low Frequency ◽  
Source Image ◽  
Medical Image Fusion

The clinical assistant diagnosis has a high requirement for the visual effect of medical images. However, the low frequency subband coefficients obtained by the NSCT decomposition are not sparse, which is not conducive to maintaining the details of the source image. To solve these problems, a medical image fusion algorithm combined with sparse representation and pulse coupling neural network is proposed. First, the source image is decomposed into low and high frequency subband coefficients by NSCT transform. Secondly, the K singular value decomposition (K-SVD) method is used to train the low frequency subband coefficients to get the overcomplete dictionary D, and the orthogonal matching pursuit (OMP) algorithm is used to sparse the low frequency subband coefficients to complete the fusion of the low frequency subband sparse coefficients. Then, the pulse coupling neural network (PCNN) is excited by the spatial frequency of the high frequency subband coefficients, and the fusion coefficients of the high frequency subband coefficients are selected according to the number of ignition times. Finally, the fusion medical image is reconstructed by NSCT inverter. The experimental results and analysis show that the algorithm of gray and color image fusion is about 34% and 10% higher than the contrast algorithm in the edge information transfer factor QAB/F index, and the performance of the fusion result is better than the existing algorithm.

Problematic plant monitoring in Hopewell Culture National Historical Park: 2008–2019

2021 ◽  
Author(s):  
Craig Young
Keyword(s):  
Plant Species ◽  
High Frequency ◽  
Low Frequency ◽  
Size Control ◽  
Medium Frequency ◽  
Rapid Spread ◽  
Exotic Invasive ◽  
Priority Species ◽  
Plant Monitoring ◽  
The Impact

Managers are challenged with the impact of problematic plants, including exotic, invasive, and pest plant species. Information on the cover and frequency of these plants is essential for developing risk-based approaches to managing these species. Based on surveys conducted in 2008, 2011, 2015, and 2019, Heartland Network staff and contractors identified a cumulative total of 51 potentially problematic plant species in Hopewell Culture National Historical Park. Of the 37 species found in 2019, we characterized 7 as very low frequency, 9 as low frequency, 17 as medium frequency, and 4 as high frequency. Of these, midpoint cover estimates of 2 medium frequency and 2 high frequency species exceeded the 10-acre threshold. Because of the number, extent, and cover of problematic plants in the park and the small park size, control efforts should focus on treating high priority species across the entire park. High priority species may include plant species capable of rapid spread, species at low population levels, and species which can effectively be controlled.

Sign in / Sign up

Export Citation Format

Share Document