- Algorithms and Data Compression
- Evolutionary Algorithms and Applications
- Genomics and Phylogenetic Studies
- Robotics and Sensor-Based Localization
- RNA and protein synthesis mechanisms
- Human Pose and Action Recognition
- Sleep and Work-Related Fatigue
- Gene expression and cancer classification
- ECG Monitoring and Analysis
- Smart Agriculture and AI
- Energy and Environment Impacts
- Machine Learning in Bioinformatics
- Indoor and Outdoor Localization Technologies
- Hybrid Renewable Energy Systems
- Spectroscopy and Chemometric Analyses
- Non-Invasive Vital Sign Monitoring
- Autonomous Vehicle Technology and Safety
- Consumer Attitudes and Food Labeling
- Human-Automation Interaction and Safety
- Hydrology and Drought Analysis
- Hydrogen Storage and Materials
- Wireless Power Transfer Systems
- Identification and Quantification in Food
- Inertial Sensor and Navigation
- Hydrological Forecasting Using AI
German Research Centre for Artificial Intelligence
2025
University of Lübeck
2023
Daffodil International University
2023
Presidency University
2023
American International University-Bangladesh
2012-2021
University of California, Riverside
2019-2020
City University
2018
North South University
2018
Islamic University of Technology
2013-2015
Recent advancements in hardware technology have spurred a surge the popularity and ubiquity of wearable sensors, opening up new applications within medical domain. This proliferation has resulted notable increase availability Time Series (TS) data characterizing behavioral or physiological information from patient, leading to initiatives toward leveraging machine learning analysis techniques. Nonetheless, complexity time required for collecting remain significant hurdles, limiting dataset...
Abstract Background Essential genes are those that critical for the survival of an organism. The prediction essential in bacteria can provide targets design novel antibiotic compounds or antimicrobial strategies. Results We propose a deep neural network predicting microbes. Our architecture called DeeplyEssential makes minimal assumptions about input data (i.e., it only uses gene primary sequence and corresponding protein sequence) to carry out thus maximizing its practical application...
Helianthus annuus, often known as sunflower, is a crop that only mildly affected by drought. The agricultural sector of the economy benefits greatly from this. However, various illnesses have imposed halt on sunflower cultivation over world. many severe diseases will affect plants if corrective measures are not taken sooner. Therefore, it negative impact yield, quantity, and quality. Diagnosing disease hand can be time-consuming difficult process. Object recognition methods use deep learning...
Predicting the class of gene expression profiles helps improve diagnosis and treatment diseases. Analysing huge data otherwise known as microarray is complicated due to its high dimensionality. Hence traditional classifiers do not perform well where number features far exceeds samples. A good set help classify dataset efficiently. Moreover, a manageable also desirable for biologist further analysis. In this paper, we have proposed linear regression–based feature selection method selecting...
To drive safely, the driver must be aware of surroundings, pay attention to road traffic, and ready adapt new circumstances. Most studies on driving safety focus detecting anomalies in behavior monitoring cognitive capabilities drivers. In our study, we proposed a classifier for basic activities car, based similar approach that could applied recognition daily life, is, using electrooculographic (EOG) signals one-dimensional convolutional neural network (1D CNN). Our achieved an accuracy 80%...
To drive safely, the driver must be aware of surroundings, pay attention to road traffic, and ready adapt new circumstances. Most studies on driving safety focus detecting anomalies in behavior monitoring cognitive capabilities drivers. In our study, we proposed a classifier for basic activities car, based similar approach that could applied recognition daily life, is, using electrooculographic (EOG) signals one-dimensional convolutional neural network (1D CNN). Our achieved an accuracy 80%...
Wearable Human Activity Recognition (HAR) is an important field of research in smart assistive technologies. Collecting the data needed to train reliable HAR classifiers complex and expensive. As a way mitigate scarcity, Time Series Data Augmentation (TSDA) techniques have emerged as promising approach for generating synthetic data. TSDA not trivial image augmentation has been relatively less investigated. In this paper, comparative study various state-of-the-art applied context wearable...
This research paper deals with a low SAR patch antenna that resonates at 2.415 GHz. can be employed to monitor the pacemaker system wirelessly, considering its performances. The main aim was operate ISM (Industrial, Scientific and Medical) band (2.4-2.48 GHz) in system, where body granted materials were used construct both ensure biocompatibility. To investigate parameters' changes between free space in-side condition, designed examined compared for conditions. key speciality of this design...
Abstract Essential genes are that critical for the survival of an organism. The prediction essential in bacteria can provide targets design novel antibiotic compounds or antimicrobial strategies. Here we propose a deep neural network (DNN) predicting microbes. Our DNN-based architecture called D eeply E ssential makes minimal assumptions about input data (i.e., it only uses gene primary sequence and corresponding protein sequence) to carry out prediction, thus maximizing its practical...
The research proposes an approach that can automatically classify real world pictures of some traditional clothing worn in Bangladesh into the predefined classes using Convolutional Neural Networks (CNN). is driven by considering growing market online shops mind. For classification purpose, we have collected images from several stores and labeled them accordingly. Our CNN model based on Google Inception model. comparison purposes tried architectures variations to see how our perform against...
The change in the normal rhythm of a human heart can lead to different cardiac arrhythmia. To monitor functionality cardiovascular system by electrocardiogram (ECG) is used reliably. Recently, accurate categorization heartbeats accepted be very concerned incident. In this paper, we applied convolution neural network (CNN) classify five classes heartbeat category and had generated classified output for information fusion towards smart monitoring (SHMS) that consist smart-phone application...
Searching for the frequent pattern within a specific genetic sequence has become much needed task in bioinformatics sector. Most recent works are based on Apriori algorithm, GSP, MacroVspan etc. techniques. However, mining can be made more efficient. In this paper, we propose two algorithms. The first one indexes unique sequences of length four using an integer value. second algorithm finds frequency patterns various lengths by searching through values instead themselves. All is done highly...
Motifs are meaningful short sequences which conserve itself during the evolution and discovery of motifs used to put DNA into their corresponding categories. Different evolutionary methods have been for motif i.e. genetic algorithm, PSO etc. In this paper we incorporated concept Linear-PSO find from sequences. However, is a slower method involves linear search discovery. So introduced function index table make faster. Before comparing target (a particle selected by in each cycle) linearly...
The involvement of a multitude parameters adds to the complexity modeling flood. However, floods are among most destructive natural disasters and therefore, flood forecasting is one key priorities hydrology. Flood goes long way minimize loss lives as well economic losses. Furthermore, proper can contribute immensely towards future risk reduction introduction necessary policies. At present, application machine learning in river analysis has dramatically increased hydrologists. In this...
Energy and economy are directly correlated. Now-a-days the development of a country is measured by energy consumption rate that country. Bangladesh's infrastructures quite small, insufficient poorly managed with per capita 220KWh only. Bangladesh steadily climbing up ladder industries shifting towards automation process. power need to act as key catalyst in helping this endeavor. The installed capacity plants around 8000MW for supporting demand 7000MW. However due failure proper fuel...
One of the fundamental goals `Digital Bangladesh ' is ensuring people's democracy, human rights, transparency, and delivering government services through maximum use modern technologies. However, Election Commission still striving to implement Electronic Voting Machines (EVM) in parliamentary general election on an extended scale due lack acceptance concerning EVM accuracy, easy accessibility, security. This paper describes implementation a low-budget securely accessible solution for Machine...
An Autonomous Vehicle (AV) is an intelligent cognitive engineering system that processes streams of observations from different onboard sources, like cameras, ultrasonic sensors, GPS units, radars, lidar's, or inertial to guide itself without human intervention. The application and research AV on multimodal disciplines are gaining attention both academia automotive companies. In this paper, we use a signal find the destination so can automatically detect route reach destination. With help it...
Recent advancement in the field of life science has caused generation massive amount genomic data. Storing such huge data require a lot memory. However, by using efficient algorithm we can optimize size dataset and save memory storage. In this paper have proposed an which uses Tetra-nucleotide RankList for optimizing storage requirement storing DNA sequences database be easily retrieved time manner. been generated testing several to confirm uniformity over all sequences. The applied on...