International Journal on Artificial Intelligence Tools

There are many dynamic real-life optimization problems in which objectives increase or decrease over time, which usually leads to variations in the dimensions of a Pareto front. Dynamic multi-objective optimization (DyMO) approaches aim to keep track of the updated Pareto front to tackle the changes which are caused by the dynamic environment. However, the current DyMO approaches do not handle dynamic environments effectively. In this study, a new hybrid dynamic two-archive evolutionary algorithm with a newly added simulated annealing and opposition-based learning strategy is proposed. The proposed method helps to preserve solutions with reasonable diversity and improve convergence by searching for promising solutions within acceptable computational time and effort. To evaluate the efficacy of the suggested method, comprehensive experiments using different multi-objective quality measures such as generational distance, and inverted generational distance have been performed on several benchmark problems with varying numbers of objectives over time. The results of the experiments show that the suggested method outperforms the strategies already in use.

Problem: In next-generation networks, users can optimize or tune their preferences with a seamless transfer of diverse access methodologies for maximizing the Quality of Service (QoS) and cost savings. In these heterogeneous wireless environments, users are prepared with several multimode wireless devices for maximizing media services through several access networks. Such networks may vary regarding energy usage, available bandwidth, technology, coverage range, monetary cost, etc. In recent days, vertical handover has attained higher performance owing to the improvements in mobility models through adopting the Fourth Generation (4G) technologies. On the other hand, these improvements are restricted to some cases, so, it does not offer support for generic mobility. Consequently, diverse strategies were implemented by considering these mobility models. However, it suffers from improper network selection, late and too-early handovers, repeated handovers, high packet loss, etc.

Aim: This paper tackles the problem of vertical handover problem in the heterogeneous network using deep learning with game theory.

Methods: The proposed model develops a non-cooperative game approach, in which all base stations compete selfishly to transmit at higher power. The overall performance in terms of throughput, handover, energy consumption, and load balancing is attained by optimizing the transmission power by the game theory. For performing this model, the required data like path loss, SINR, data rate, load, etc are generated by the deep learning called Recurrent Neural Network (RNN).

Results: From the simulation findings, the handoff probability of the recommended RNN+Game Theory is correspondingly secured at 6.9%, 22.6%, and 8.2% superior to TOPSIS, ABC-PSO, and game theory when taking the time like 5 secs for user velocity as 30 km/h.

Conclusion: Results show that the proposed game theoretical approach with deep learning provides a throughput enhancement while reducing the power consumption, in addition, to minimizing the unnecessary handover and balancing the load between base stations.

Diabetic retinopathy is a predominant vision-threatening disease affecting working-aged people specifically. Timely diagnosis through early detection and prevention helps to reduce the risk of severe vision loss. Computer-aided diagnosis in retinal image analysis through Machine Learning techniques will help medical professionals perform their analysis better. Automated image processing through Convolutional Neural Networks has proven to be a promising technique, mainly in medical image segmentation. Convolutional Neural Network techniques like 3D CNN, Deep CNN and architectures like U-Net, V-Net, SegNet, and DeepMedic have outperformed medical image analysis results. However, the baseline CNN architectures struggle to retain high-quality information at the output and thus affect the performance. This work focuses on addressing the issue of translational variance and overfitting scenarios of U-Net architecture by experimenting with adaptable window sizes, pretrained weights and linear interpolation technique. A novel U-Net based architecture called RetU-Net that segments abnormal retinopathy lesion structures by retaining higher-level feature is proposed. The experiments are conducted using Indian Diabetic Retinopathy Image Dataset provided in “Diabetic Retinopathy: Segmentation and Grading Challenge” initially. The results have been compared with other state-of-art CNN architectures. Further evaluation is carried out on public datasets: STARE and DRIVE, and the performance is compared with U-Net based architectures used for retinal image segmentation. The proposed approach is efficient in terms of precision, sensitivity, specificity and accuracy. It received accuracy scores of 0.9612, 0.9712 while experimenting with STARE and DRIVE datasets respectively.

Pedestrian detection is one of the most challenging research areas in computer vision. Compared to traditional hand-crafted methods, convolutional neural networks (CNNs) have superior detection results. The single-stage detection networks, particularly the state-of-the-art You Only Look Once (YOLO) network, have attained a satisfactory performance without compromising the computation speed in object detection. YOLO framework can be leveraged in pedestrian detection as well. In this work, we propose an improved YOLOv2, called DSM-IDM-YOLO. The proposed model uses a modified DarkNet19 integrated with three new modules, two depth-wise separable convolution modules and one inception depth-wise convolution module, leading to a comprehensive feature of an object in the image. The modules are integrated on top of features from multiple levels of the network. The proposed framework is computationally less expensive owing to its convolution design and a moderate number of layers. It aims to improve performance with minimal computational overhead. The proposed method is compared with state-of-the-art detection methods, i.e., Faster R-CNN, YOLOv2, YOLOv3, YOLOv4-tiny and Single Shot Multibox Detector (SSD). The performance results attest that the proposed method has effectively improved the detection. Three benchmark pedestrian datasets are used for experimental analysis: INRIA, PASCAL VOC 2012 and Caltech.

Fatigue detection of workers is an important factor in construction site monitoring. Nowadays, worker exhaustion on construction sites causes tiredness and drowsiness. The prediction of mental exhaustion is critical because the job has increased over the years. Accurate fatigue detection is important for analyzing the stress level of work on construction sites. However, recording worker activities and detecting fatigue is critical for site supervisors. Over the last century, there has been an increasing trend towards vision-based action recognition. The identification of worker activities in far-field surveillance video has received good attention. Henceforth, this research proposed a hierarchical statistical approach for detecting worker activity in far-field security footage. In this paper, extension based equilibrium with capsule autoencoder network is proposed for fatigue detection. To improve the prediction performance, video data is collected for monitoring. Initially, the video dataset is converted into frames, and these frames are pre-processed using normalization. Afterwards, statistical feature extraction techniques are extracted from the normalized frames to improve the prediction performance. It helps to identify the symptoms of fatigue detection. The persons with the symptoms are identified from the extraction. Feature selection is performed using hybrid Battle Royale Optimization and Particle Swarm Optimization algorithms (hybrid BRO-PSO). Finally, based on the selected features, fatigue classification is performed using extended equilibrium optimization with Capsule Autoencoder (EECAEN). The performance of the proposed scheme is compared with different existing approaches. The performance of fatigue detection is analyzed with real time video dataset. The performance of the proposed approach is compared with different existing techniques such as Convolution Neural Network (CNN), Long Short-Term Memory (LSTM), and Recurrent Neural Network (RNN). The proposed method outperforms the existing approaches in terms of F1-score, accuracy, recall, precision, specificity, and sensitivity.

For predicting missile’s interception point, the current guidance law based on neural networks avoids to model the strong nonlinear motion of a missile and simultaneously improve the anti-jamming ability of the guidance law. Although the advantages of solving the predicted intercept point problem based on neural networks are obvious, the difficulty in obtaining the target missile information still exists. In this work, we propose a dead-reckoning navigation guidance (DRNG) law. First, a neural network-based collaborative forecast scheme is proposed and utilize the advantages of different neural networks to greatly reduce the difficulty in acquiring the target information. Second, we construct an approximate realistic aerodynamic characteristics environment to simulate the motion parameters of missiles and targets. We also introduce real-time error correction for increasing the prediction accuracy of the network and improve the robustness of the proposed DRNG by using the model self-update algorithm. Finally, through a large number of simulation experiments, results show that the proposed DRNG completes the interception task in a noisy environment, when only the position parameters of a target missile are known. Moreover, it has a more optimized ballistic trajectory as compared with the traditional guidance law.

Computer Vision (CV) has become an essential field in Artificial Intelligence applications. Object detection and recognition (ODR) is one of the fundamental tasks of computer vision implementations. However, developing an efficient ODR model is still a significant problem. The model’s execution time and speed are the most critical features during the inference or detection and recognition process, which need to be improved using the latest object detection architectures. In this paper, the handwritten detection and recognition (HDR) model is developed based on previously known algorithms with their efficiency, such as Faster R-CNN and YOLOv4 in the first hand. On the other hand, two new models capable of detecting and recognizing handwritten digits using the latest ODR algorithm are proposed, one based on the latest YOLO family architecture (YOLOv5-HDR) with high speed and accuracy and the other using the transformers architecture (DETR). To the best of our knowledge, this is the first study to achieve a details comparison between YOLOv5 and transformers-based models in handwritten digit detection. Finally, the detailed performance analysis achieved by the paper proves that the YOLOv4-based model achieved the testing inference 13% faster than Faster R-CNN. However, the proposed YOLOv5-based model outperformed the YOLOv4 and the transformers-based one as it increased the testing execution time 25% faster than the YOLOv4, three times faster than the DETR model. A further adversarial attack test has been conducted to ensure the robust performance of the proposed model. Furthermore, numerical experiment results and their analyses demonstrate the robustness and effectiveness of the proposed YOLOv5-based model being the most stable for handwritten digit detection and recognition tasks.

The holistic understanding of the information contained in technical documents depends on the understanding of the document’s individual modalities. These modalities are tables, graphics, diagrams, formulas, etc. and each of them is a standalone research topic that requires a different way of processing and understanding. These modalities, processed and combined with the document text, can introduce new techniques for visual question answering in documents. Thus, in this paper we present Pinakas, a methodology for automatic analysis of the internal tabular information that appears in technical documents and its modeling to stochastic Petri-net graphs. We focus only on tables that strictly abide by the IEEE format rules. The methodology presented here is divided into the following steps: (1) table detection, (2) table recognition, (3) table understanding. Qualitative results of Pinakas are demonstrated as proof of concept for the accurate extraction of information from three different types of tables.

The K-variant is a multi-variant architecture to enhance the security of the time-bounded mission and safety-critical systems. Variants in the K-variant architecture are generated by controlled source program transformations. Previous experimental studies showed that the K-variant architecture might improve the security of systems against memory exploitation attacks. In order to estimate the survivability of K-variant systems, simulation techniques are utilized. However, these techniques are slow and may not be practical for the design of K-variant systems. Therefore, fast and highly accurate estimations of the survivability of K-variant systems are necessary for developers. The neural networks may allow quick and accurate estimation of the survivability of K-variant systems. The developed neural network-based tool can make quick and precise estimations of the survivability of K-variant systems under different conditions. In this paper, the accuracy of the neural network-based tool is investigated in an experimental study. The neural network-based tool estimations are compared with a K-variant attack emulator in three programs for up to ten variant systems under four attack types and three attack durations. The experimental study demonstrates that the neural network-based tool makes fast and accurate estimations of the survivability of K-variant systems under all the conditions investigated.

Classification of tumors from MRI plays very important role for diagnosing various diseases. But, it consumes an enormous amount of time for classification. Due to the similar structure of anomalous and typical tissues in the brain, it is difficult to complete the detection process successfully. Many researchers have developed new methods for detection and classification of tumors. But most of them failed at some point due to these limitations. Therefore in our work, we introduced a new machine learning algorithm for detection and classification of tumors. In addition to this, an intellectual segmentation technique known as Improved Binomial Thresholding technique is also introduced in this paper. This newly developed approach is used to differentiate the normal and abnormal slices from brain MRI. We can extract different features from the segmented image. The extracted features may be Wavelet Transform based (WT) or Scattering Wavelet Transform based (SWT). Feature selection process is achieved using a hybrid algorithm known as CS-FS (Hybrid Cuckoo Search with Fish Swarm) to minimize the dimension of extracted features. Finally, feature classification process is performed using Sparse Bayesian Extreme Learning Machine (SBELM) classifier. The proposed method is executed with the help of BRATS (Brain Tumor Image Segmentation) 2015 dataset. The result of the proposed method is evaluated on different parameters like accuracy, specificity, and sensitivity. The values of these parameters are obtained by computing four factors such as TP, TN, FN and FP. The final evaluation results showed that, our proposed SBELM classifier has attained 97.2% accuracy, which is better than the other existing method like state-of-the-art method.