In the real world the intensity inhomogeneity occurs presenting considerable challenges in the field of segmentation of images. In the arena of computer vision and processing of images, the method ...

The Covid-19 disease caused by the novel Corona Virus (SARS-2) spread like a wildfire and scientists across the whole world have been trying to find a cure for the disease and as such resorted to all methods available. The tool of artificial intelligence (AI) and data science has proven very useful in this regard for rapid drug invention and development. In this paper, tending along the same line, four different deep neural networks (DNNs) based models (bi-directional long short-term memory (LSTM) with attention, constrained graph variational autoencoders (CGVAE), edge memory neural network (EENN) and connectivity map (CMAP) based DNN have been proposed for usage in drug Invention of highly effective lead molecules for the disease COVID-19. The models have been evaluated and performed well with the highest performance given by the bi-directional LSTM model with validity of 98.7%, uniqueness of 99.8% and originality of 97.4%.

Generally, mutation testing has been considered effective to test the adequacy of the test suite over a set of artificial faults. These faults are created by applying different mutation operators o...

This paper proposes a new hybrid, optimal cost based routing approach namely HGAB3C implemented for routing in wireless mesh networks (WMNs). We used this proactive routing approach to find the optimal paths in dynamically created WMN of client nodes. The performance of proposed approach was evaluated with different architectural scenarios from 1000 nodes to 10,000 nodes WMNs with a step increase of 1000 nodes. The route costs were evaluated using integrated link cost (ILC) measure which is a function of jitter, network delay and throughput. Performance of this proposed approach was compared with seven other soft computing based approaches and two hard computing based algorithms. The proposed approach outperformed all the other 9 competing approaches. The proposed algorithm can also be used effectively in isolation wards of COVID-19 hospitals and quarantine centers where robots can be networked for communications between patients and doctors.

Present study proposes an approach for the estimation of the distinctiveness of human footprints under the machine learning environment. In this system, a sum of 880 raw footprints have been segmented to get the 21 features for ensemble learning. All the features have been analysed for computation of minimum, mean grey value, median, maximum, standard deviation, kurtosis, and skewness for footprint dataset. The G-means clustering offers centroid information of footprint features. A set of 10 ensembles has analysed for surrogate footprint attributes. Ten anomaly models were created for anomaly scores among these features. The association of features gives the uniqueness of the human footprints for personal identification through fuzzy rules for every set of ensembles. As a consequence, centroid, ensembles, anomaly, and affiliation proved the individuality of human footprints.

Human face gender identification is increasingly gaining popularity because of exponential interest in ubiquitous and pervasive computing. The computing embedded in environment can feel the person's presence and as per person being male or female, may induce certain decisions with help of ubiquitous computing devices to make environment suited to person. The challenge of detecting a face is male or female is very trivial due to similarity of features. This work presents use of Thepade's Sorted Block Truncation Coding N-ary (TSBTC N-ary) for face feature extraction and further deploys machine learning classifiers to identify face as male or female. In proposed face gender identification, TSBTC N-ary is explored with six combinations (from two-ary to seven-ary) for face feature extraction with fourteen machine learning classifiers giving 96 variations; tested using Faces94-dataset. Classification accuracy is used as performance measure. Overall Random Forest gives best performance and TSBTC-7ary outperforms other feature extraction variations.

Cholera is one of the rapidly spreading waterborne diseases which are caused by bacteria named Vibrio Cholerae. As population grows, so as the data related to patients, doctors and health staff. So far, different types of machine learning models have been proposed for classification of Cholera infection. However, the majority of these suffers from pre-mature convergence and stuck in local optimal issues. In this paper, ANFIS-GA based forecasting model for the prediction of Cholera virus has been proposed. In the proposed model, non-dominated sorting genetic algorithm (NSGA) is used to tune hyper-parameters of ANFIS. The comparisons are done among the designed NSGA-ANFIS and the existing models on the benchmark Cholera dataset. Performance analysis illustrates that the designed NSGA-ANFIS model performs significantly better than the existing models such as ANFIS, PSO-ANFIS and GA-ANFIS in terms of accuracy, sensitivity, kappa statistics, specificity and F-measure, as 99.2%, 99.04%, 99.11%, 99.49% and 98.85%, respectively.