Volume 11 | Issue 8 |

 Abstract

Agriculture is an important source of livelihood and Indian economy depends on agricultural production. It is important to detect the plant leaf diseases at early stage to increase the crop yield and profit. Image processing technique is used to detect the leaf diseases accurately since naked eye observation of the diseases does not provide accurate result all the time especially during the early stage. It was done in five phases which are image acquisition, preprocessing of the acquired image, feature extraction, classification of the diseases and displaying the result. This paper provides a detailed survey about classification of the agricultural diseases by using Support Vector Machine classifier.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Accurate,Acquisition, Segmentation, Support Vector Machine.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In this research, we show how machine learning may be used to make wired and wireless TCP networks more resilient to congestion. Since TCP reacts the same way to losses caused by congestion as it does to losses caused by link faults, it is sub- optimal in hybrid wired/wireless networks. So, we propose simulating various network topologies and then using machine learning to create a loss classifier automatically. Several machine learning methods are compared, and decision tree boosting is found to be the most effective. In comparison to ad hoc classifiers from the networking literature, it performs better.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Congestion Management in Wireless Networks: A Machine Learning Approach

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Cheque Truncation System is utilized for clearing cheques. Because of the restrictions of this semi-manual procedure, it can take up to three business days for a national interbank check to clear in Sri Lanka. In light of these drawbacks, it is evident that both cheque users and large financial institutions want a more streamlined and protected system, one that can clear a check in under 24 hours while maintaining the system's security and privacy. To address these concerns, this study presents an automated approach that is within the reach of every financial institution in Sri Lanka. To provide its customers with the faster cheque clearance, the suggested system is built on the blockchain, and all banks that are interested in that structure must link to it. A comprehensive framework was proposed, with answers split into four distinct phases: (i) the traditional paper check clearing process, (ii) the digital cheque producing and cleared process, (iii) the detection and prevention of fraudulent cheque transactions, and (iv) the safeguarding of cheque transactions. The main programming languages and frameworks utilized to build the system were Python and Flutter, while the main distributed ledger used was Ethereum. Because Ethereum adds extra security, the suggested system may become very large very quickly. Both the consumer and the bank benefit from the method, since the clearing of cheques is made easier and faster while also being made more secure. Paper check fraud detection is made easier and more accurate as a consequence.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Online Digital Cheque Clearance and Verification System Using Block Chain

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Due to the broad availability of cameras, photo taking has been becoming increasingly common in recent years. Images play a crucial role in our everyday lives because to the abundance of information they hold, and it is frequently necessary to modify images to get new insights. There are many options for enhancing photos, but they are also regularly exploited to create fake photos that distribute false information. This is really worrisome since it increases the incidence and severity of picture forgeries. Over time, several conventional methods have been developed to identify picture frauds. The topic of visual forgery detection has been inspired by convolutional neural networks (CNNs), which have gained a lot of attention in recent years. While CNN-based picture forgery detection methods do exist, they typically only identify one sort of fraud (such as image merging or copy-move) at a time. Therefore, a method is needed that can effectively andefficiently detect the existence of unnoticed forgeries in a picture. In this research, we provide a powerful deep learning-based approach for spotting fakes in the setting of double compression of pictures. We train our representation on the disparity between the uncompressed and compressed variants of a picture. The suggested model requires fewer computational resources and has been shown to outperform the current best methods. Positively, experimental findings show a validation validity rate of 92.23 percent overall.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Using a Concatenation of Low-Complexity Deep Learning Models, We Can Spot Fake Images

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

There are currently no reliable ways to determine if it will rain today. Estimates made by the meteorological agency themselves may turn out to be inaccurate. In this post, we will learn how to create a machine-learning model that can forecast whether or not it will rain today based on various atmospheric variables. Because machine learning models typically outperform human beings at the previously known task--predicting rainfall--this issue is connected to rainfall prediction using machine learning.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

RAINFALL PREDICTION BASED ON MACHINE LEARNING

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

It's no wonder that criminals would want to take advantage of the financial system by flooding it with counterfeit notes that seem very comparable to the genuine thing, given that bank cash is our country's greatest asset. When demonetization takes place, counterfeit banknotes flood the economy. A genuine banknote and a counterfeit one have numerous characteristics, making it difficult, if not impossible, to tell them apart without special equipment or training. It is a challenging task to distinguish real banknotes from fake ones. Therefore, a completely automated system is essential, and it must be available through tellers at banks and ATMs. Because counterfeit banknotes may now be made with such high quality, it is crucial that a reliable algorithm be created to evaluate whether or not a

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

The Use of Machine Learning to Spot Counter feit Currency

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Evidence management is highly valued in the field of forensic science. In order to solve crimes and bring perpetrators to justice, evidence gathered at the site of the crime is essential. Therefore, it is essential to keep these goods protected from any kind of manipulation. Chain of custody is a method used to ensure that evidence is kept secure. If the chain if custody is broken, the evidence cannot be used in court, which might lead to an automatic dismissal of the proceeding. Because the standard practice for handling forensic evidence is so wasteful, moving toward a digital system is crucial. Blockchains are public, digitally distributed ledgers comprising transactions that have been signed cryptographically and divided into blocks. The Linux Foundation created Hyperledger Fabric, a blockchain platform designed specifically for enterprise use cases. Based on the concepts of Hyperledger Fabric, this study aimed to provide a framework for implementing Blockchain Technology into the criminal records administration system while maintaining the Chain of Custody.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

AN IMPLEMENTATION OF BLOCKCHAIN TECHNOLOGY IN FORENSIC EVIDENCE MANAGEMENT

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Among many types of web-based applications, real-time streaming footage has become the standard broadcast standard. In this paper, we investigate Amazon Prime Video, which is an online video-on-demand service that allows us to view sporting activities, motion pictures, and other events. Each component of the video that streams online is then further broken into various bit rate variants. The quantity of bits sent in a specific amount of time is known as the rate of transmission for videos. Utilizing the wireless network analysis programme Wire shark, we quantify the bit rates modulated throughput of Amazon Prime Video Streaming. The rate of modulation depends on a pair of factors, like traffic rate and video resolution; the latter is used to gauge a video's capacity and network capacity, which takes into account the higher bandwidth requirement, the better the quality must be.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Wireshark, Amazon Prime, Bandwidth, Resolution, Streaming

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Satellite photos are widelyusedinfieldsincludingemergencymanagement,security,andenvironmentalmonitoring.Thesegoalscan'tbeachievedwithout the help of humans and the ability toproperlyidentifyobjects.Withsomanypossiblesearchspaces and so few analystson hand, automation is essential. However,owingtotheirfocusonaccuracyandprecision,conventionaltechniquestoidentifyitemsandcategorizationareconstrainedintheircapacitytodeliverasolution.Automatingthesestepsusingsupervised neural class ML algorithms hasshownsomesuccess.Thereissomeevidence that convolutional neural networks,akindofartificialneuralnetwork,mayenhancebothpictureidentificationandunderstanding. In this case, we use them tolearnhowtoidentifyartificialfeaturesinhigh-resolution,multispectralsatelliteimagery.Weprovideadeeplearningapproachtoclassifyingfeaturesor

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Classification of Satellite Images Using Deep Learning

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Congestioninbigcitiesisbecomingmoreproblematicas urban populations & automobileownership both grow. Everyone suffers inconvenient delays when traffic is backed up. motorists,but it does increase fuel consumption and pollution. Despite its apparent global prevalence, bigurbancentersare especially at risk. Therefore, it is becoming more vital to calculate trafficdensityinrealtimeinorder to optimize signal timing and manage traffic flow. The trafficcontroller plays a crucial role in the smooth flow of traffic. Therefore, there is an urgent need forenhanced traffic management to fulfill the expanding demands of the public. To measure trafficlevels, our system would utilize artificial intelligence and processing of images to examine livefeeds via cameras installed at crossings. The algorithm for adjusting the timing of traffic lights inresponse to traffic volumes is also highlighted in an effort to make travel more efficient forpassengers and reduce pollution.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

MANAGINGS TOP AND GO TRAFFIC WITH AI

  License

Creative Commons Attribution 4.0 and The Open Definition