This study presents a hybrid Generative Adversarial Network (GAN) and Support Vector Machine (SVM) model for accurately categorizing Raspberry Cane Disease (RCD) into four severity levels. Leveraging a dataset of 7,00...
详细信息
This study presents a hybrid Generative Adversarial Network (GAN) and Support Vector Machine (SVM) model for accurately categorizing Raspberry Cane Disease (RCD) into four severity levels. Leveraging a dataset of 7,00...
详细信息
ISBN:
(数字)9798331529635
ISBN:
(纸本)9798331529642
This study presents a hybrid Generative Adversarial Network (GAN) and Support Vector Machine (SVM) model for accurately categorizing Raspberry Cane Disease (RCD) into four severity levels. Leveraging a dataset of 7,000 high-resolution images, the model achieves 98.8% accuracy. GANs generate synthetic images to address data scarcity, while SVMs ensure precise multi-class classification. The results demonstrate the model's effectiveness in disease severity assessment, highlighting its potential for broader agricultural applications.
The proposed work looks into differences between types of Kohlrabi common in India using special systems called Convolutional Neural Networks (CNN) with Genetic Algorithms (GA). A set of 4000 high-quality pictures was...
详细信息
ISBN:
(数字)9798331512248
ISBN:
(纸本)9798331512255
The proposed work looks into differences between types of Kohlrabi common in India using special systems called Convolutional Neural Networks (CNN) with Genetic Algorithms (GA). A set of 4000 high-quality pictures was carefully gathered and readied. These included Purple Vienna, White Vienna, and Grand Duke types. The CNN design was made to find small differences in pictures. It used convolutional, pooling, and fully connected parts for a strong grouping of things. The CNN model was good at telling apart different kinds of kohlrabi. It got it right 96.12 % of the time based on how they looked and what color they were! The test results showed that the model was very accurate across all categories, showing it is reliable in spotting small differences. Comparing it to other ways of doing things, found that the CNN model worked better. This made sure it is a strong and trustworthy choice for telling different types of Kohlrabi apart. This study shows that machine learning methods, especially CNNs can be good for farming tasks. It gives hope to finding the right type of plants better in farm work and separate markets for sales. Also, these techniques help improve breeding programs. The good use of CNNs in correctly grouping Kohlrabi types shows better ways to grow crops and can help with making food long-lasting. It also leads people closer to understanding what different plants look like, helping those who work on farms make smarter choices.
For the practical deployment of optical packet switching (OPS) system, realizing all-optical buffering is a critical issue. In this paper, we present a hybrid labelling scheme based on optical code-division multiplexi...
详细信息
For the practical deployment of optical packet switching (OPS) system, realizing all-optical buffering is a critical issue. In this paper, we present a hybrid labelling scheme based on optical code-division multiplexing (OCDM) over generalized multi-protocol label switching (GMPLS) network. The virtual optical memory is implemented by assigning queuing packets with different coded-WDM labels. The proposed hybrid labels combine the advantages of wavelength-division multiplexing (WDM) with spectral-amplitude-coding (SAC) OCDM. Due to the cyclic-shift property of arrayed-waveguide grating (AWG) and maximal-length sequence (M-sequence) codes, blocks of label generating and processing in node organization can be built in low complexity. Furthermore, utilizing hybrid coding technique in the buffering node increases the efficiency of label generating and processing. Derivation of the probability of packet missing is done to quantify the buffering performance of virtual optical memory. Simulation results show that the hybrid-coded labels, as compared to the conventional OCDM labels, are able to buffer the data packets with a relatively low probability of packet missing under similar label length.
Recently, we proposed an innovate crossbar switch named Contention-tolerant Crossbar Switches (CTC(N)). Different with conventional crossbar switches which resolve output contentions by building conflict-free connecti...
详细信息
In this paper, we study how to collect n balls moving with constant velocities in the Euclidean plane by k robots moving on straight track-lines through the origin. Since all the balls might not be caught by robots, d...
详细信息
In this paper, we study how to collect n balls moving with constant velocities in the Euclidean plane by k robots moving on straight track-lines through the origin. Since all the balls might not be caught by robots, d...
详细信息
In this paper, we study how to collect n balls moving with constant velocities in the Euclidean plane by k robots moving on straight track-lines through the origin. Since all the balls might not be caught by robots, differently from Moving-Target TSP, we consider the following 3 problems in various situations: (i) deciding if k robots can collect all n balls, (ii) maximizing the number of the balls collected by k robots, and (iii) minimizing the number of the robots to collect all n balls. The situations considered here contain the cases in which track-lines are given (or not), and track-lines are identical (or not). For all problems and situations, we provide polynomial time algorithms or proofs of intractability, which clarify the tractability-intractability frontier in the ball collecting problems in the Euclidean plane.
This paper examines the impact of RED on two versions of TCP - traditional TCP Reno and a newly proposed variant, TCP Veno - over 802.11b WLAN. TCP Reno was originally designed for wired networks where packet losses a...
详细信息
This paper examines the impact of RED on two versions of TCP - traditional TCP Reno and a newly proposed variant, TCP Veno - over 802.11b WLAN. TCP Reno was originally designed for wired networks where packet losses are primarily due to network congestion. This assumption is not always true in wireless networks, in which packet losses can be due to transmission errors on the noisy wireless link. TCP Veno refines the algorithms in Reno by distinguishing between noncongestive and congestive states, and avoids the unnecessary reduction of TCP congestion window when packet losses are not due to congestion. Our results show that TCP Veno can achieve up to 30% more throughput than TCP Reno when link quality is poor. Our results also show that TCP Veno is compatible with RED. In addition, although RED does not help to further improve the throughput in Veno, it can improve fairness among co-existing TCP flows.
暂无评论