arXiv subjects

A. Sufian

Publications and source records attributed to A. Sufian.

7 recordsLinked to original sources

arXiv2020

BDNet: Bengali Handwritten Numeral Digit Recognition based on Densely connected Convolutional Neural Networks

Images of handwritten digits are different from natural images as the orientation of a digit, as well as similarity of features of different digits, makes confusion. On the other hand, deep convolutional neural networks are achieving huge success in computer vision problems, especially in image classification. BDNet is a densely connected deep convolutional neural network model used to classify (recognize) Bengali handwritten numeral digits. It is end-to-end trained using ISI Bengali handwritten numeral dataset. During training, untraditional data preprocessing and augmentation techniques are used so that the trained model works on a different dataset. The model has achieved the test accuracy of 99.775%(baseline was 99.40%) on the test dataset of ISI Bengali handwritten numerals. So, the BDNet model gives 62.5% error reduction compared to previous state-of-the-art models. Here we have also created a dataset of 1000 images of Bengali handwritten numerals to test the trained model, and it giving promising results. Codes, trained model and our own dataset are available at: {https://github.com/Sufianlab/BDNet}.

cs.CV↗

arXiv2019

A Review of Object Detection Models based on Convolutional Neural Network

Convolutional Neural Network (CNN) has become the state-of-the-art for object detection in image task. In this chapter, we have explained different state-of-the-art CNN based object detection models. We have made this review with categorization those detection models according to two different approaches: two-stage approach and one-stage approach. Through this chapter, it has shown advancements in object detection models from R-CNN to latest RefineDet. It has also discussed the model description and training details of each model. Here, we have also drawn a comparison among those models.

cs.CV↗

arXiv2019

Data Load Balancing In Mobile Ad Hoc Network Using Fuzzy Logic (DBMF)

Volume and movement of data rapidly increasing in every type of data communications and networking, and ad hoc networks are not spared from these challenges. Traditional Multipath routing protocols in Mobile Ad-hoc Networks (MANETs) did not focus on data load distribution and balancing as much as required. In this scheme, we have proposed data load distribution and balancing through multiple paths simultaneously. We have considered three important parameters of ad hoc network those are: mobility of node, the energy of node and packet drop rate at a node. This scheme combines these three metrics using fuzzy logic to get the decisive parameter. We have shown improvement of this scheme over similar kind of protocols in NS-2 network simulator.

cs.NI↗

arXiv2019

Cheat-Proof Communication through Cluster Head (C3H) in Mobile Ad Hoc Network

The mobile ad hoc network (MANET) is a wireless network based on a group of mobile nodes without any centralised infrastructure. In civilian data communication, all nodes cannot be homogeneous-type and not do a specific data communication. Therefore, node co-operation and cheat-proof are essential characteristics for successfully running MANETs in civilian data communication. Denial of service and malicious behaviour of the node are the main concerns in securing successful communication in MANETs. This scheme proposed a generic solution to preventing malicious behaviour of the node by the cluster head through the single hop node clustering strategy.

cs.NI↗

arXiv2019

Fuzzy Route Switching For Energy Preservation (FEP) in Ad Hoc Networks

Nodes in ad hoc networks have limited battery power. Hence they require an energy-efficient technique to improve average network performance. Maintaining energy-efficiency in ad hoc networks is really challenging because highest energy efficiency is achieved if all the nodes are always switched off and energy-efficiency will be minimum if all the nodes are fully operational i.e. always turned-on. Energy preservation requires redirection of data packets through some other routes having good performance. This improves the data packet delivery ratio and the number of alive nodes decreasing the cost of messages.

cs.NI↗

arXiv2019

Fuzzy-Controlled Scheduling of Route-Request Packets (FSRR) in Mobile Ad Hoc Networks

In ad hoc networks, the scheduling of route-request packets should be different from that of message packets, because during transmission of message packets the location of the destination is known whereas in route discovery this is not known in most of the cases. The router has to depend upon the last known location, if any, of the destination to determine the center and radius of the circle that embeds all possible current position of the destination. Route-request packets generated from the source are directed towards this circle i.e., directional route discovery can be applied. Otherwise, when no earlier location of the destination is known the route-requested has to be broadcast in the whole network consuming a significant amount of time than directional route discovery. The present article proposes fuzzy controlled scheduling of route-request packets in particular that greatly reduces the average delay in route discovery in ad hoc networks.

cs.NI↗

arXiv2019

Advancements in Image Classification using Convolutional Neural Network

Convolutional Neural Network (CNN) is the state-of-the-art for image classification task. Here we have briefly discussed different components of CNN. In this paper, We have explained different CNN architectures for image classification. Through this paper, we have shown advancements in CNN from LeNet-5 to latest SENet model. We have discussed the model description and training details of each model. We have also drawn a comparison among those models.

cs.CV↗