1
|
Author Correction: Markov chain-based impact analysis of the pandemic Covid-19 outbreak on global primary energy consumption mix. Sci Rep 2024; 14:10444. [PMID: 38714833 PMCID: PMC11076614 DOI: 10.1038/s41598-024-61288-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 05/10/2024] Open
|
2
|
Advanced machine learning approach for DoS attack resilience in internet of vehicles security. Heliyon 2024; 10:e28844. [PMID: 38681562 PMCID: PMC11053265 DOI: 10.1016/j.heliyon.2024.e28844] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/10/2023] [Revised: 03/18/2024] [Accepted: 03/26/2024] [Indexed: 05/01/2024] Open
Abstract
Recent years have witnessed security as a great concern in vehicular networks (VANET). Particularly, Denial of Service (DoS) and Distributed Denial of Service (DDoS) attacks can jeopardize the network by broadcasting a storm of packets. Correspondingly, the network resources are jammed with malicious traffic. In this connection, the existing research presented various techniques to cope with DoS and DDoS attacks. Different from those traditional approaches, this study proposes an Intelligent Intrusion Detection System (IDS) by leveraging Machine Learning (ML). The proposed IDS utilizes a publicly available dataset on the application layer for mitigating DDoS attacks. The designed ML-based IDS relies on combining both the Random Projection (RP) and Randomized Matrix Factorization (RMF) methods to achieve the best results for enhancing the detection capabilities of the IDS. This amalgamation enhances the system's detection capabilities by extracting and analyzing meaningful features from network traffic data. Experimental validation of our approach involves a comprehensive evaluation of various ML models, including Extra Tree Classifier (ETC), Logistic Regression (LR), and Random Forest (RF). Remarkably, the combined accuracy of these models yields an average system accuracy of 0.98, surpassing existing methods. Unlike conventional approaches, our proposed IDS excels in efficiency and exhibits notable performance in detecting DoS and DDoS attacks in VANET. This proficiency ensures the integrity and safety of vehicle communications. Thus, our research substantially contributes to the vehicular network security field. The presented findings establish a foundation for future advancements in securing connected vehicles.
Collapse
|
3
|
Security risk models against attacks in smart grid using big data and artificial intelligence. PeerJ Comput Sci 2024; 10:e1840. [PMID: 38686008 PMCID: PMC11057646 DOI: 10.7717/peerj-cs.1840] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/26/2023] [Accepted: 01/09/2024] [Indexed: 05/02/2024]
Abstract
The need to update the electrical infrastructure led directly to the idea of smart grids (SG). Modern security technologies are almost perfect for detecting and preventing numerous attacks on the smart grid. They are unable to meet the challenging cyber security standards, nevertheless. We need many methods and techniques to effectively defend against cyber threats. Therefore, a more flexible approach is required to assess data sets and identify hidden risks. This is possible for vast amounts of data due to recent developments in artificial intelligence, machine learning, and deep learning. Due to adaptable base behavior models, machine learning can recognize new and unexpected attacks. Security will be significantly improved by combining new and previously released data sets with machine learning and predictive analytics. Artificial Intelligence (AI) and big data are used to learn more about the current situation and potential solutions for cybersecurity issues with smart grids. This article focuses on different types of attacks on the smart grid. Furthermore, it also focuses on the different challenges of AI in the smart grid. It also focuses on using big data in smart grids and other applications like healthcare. Finally, a solution to smart grid security issues using artificial intelligence and big data methods is discussed. In the end, some possible future directions are also discussed in this article. Researchers and graduate students are the audience of our article.
Collapse
|
4
|
Markov chain-based impact analysis of the pandemic Covid-19 outbreak on global primary energy consumption mix. Sci Rep 2024; 14:9449. [PMID: 38658780 PMCID: PMC11043445 DOI: 10.1038/s41598-024-60125-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/29/2023] [Accepted: 04/18/2024] [Indexed: 04/26/2024] Open
Abstract
The historic evolution of global primary energy consumption (GPEC) mix, comprising of fossil (liquid petroleum, gaseous and coal fuels) and non-fossil (nuclear, hydro and other renewables) energy sources while highlighting the impact of the novel corona virus 2019 pandemic outbreak, has been examined through this study. GPEC data of 2005-2021 has been taken from the annually published reports by British Petroleum. The equilibrium state, a property of the classical predictive modeling based on Markov chain, is employed as an investigative tool. The pandemic outbreak has proved to be a blessing in disguise for global energy sector through, at least temporarily, reducing the burden on environment in terms of reducing demand for fossil energy sources. Some significant long term impacts of the pandemic occurred in second and third years (2021 and 2022) after its outbreak in 2019 rather than in first year (2020) like the penetration of other energy sources along with hydro and renewable ones in GPEC. Novelty of this research lies within the application of the equilibrium state feature of compositional Markov chain based prediction upon GPEC mix. The analysis into the past trends suggests the advancement towards a better global energy future comprising of cleaner fossil resources (mainly natural gas), along with nuclear, hydro and renewable ones in the long run.
Collapse
|
5
|
Emotion detection from handwriting and drawing samples using an attention-based transformer model. PeerJ Comput Sci 2024; 10:e1887. [PMID: 38660197 PMCID: PMC11041987 DOI: 10.7717/peerj-cs.1887] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/10/2023] [Accepted: 01/29/2024] [Indexed: 04/26/2024]
Abstract
Emotion detection (ED) involves the identification and understanding of an individual's emotional state through various cues such as facial expressions, voice tones, physiological changes, and behavioral patterns. In this context, behavioral analysis is employed to observe actions and behaviors for emotional interpretation. This work specifically employs behavioral metrics like drawing and handwriting to determine a person's emotional state, recognizing these actions as physical functions integrating motor and cognitive processes. The study proposes an attention-based transformer model as an innovative approach to identify emotions from handwriting and drawing samples, thereby advancing the capabilities of ED into the domains of fine motor skills and artistic expression. The initial data obtained provides a set of points that correspond to the handwriting or drawing strokes. Each stroke point is subsequently delivered to the attention-based transformer model, which embeds it into a high-dimensional vector space. The model builds a prediction about the emotional state of the person who generated the sample by integrating the most important components and patterns in the input sequence using self-attentional processes. The proposed approach possesses a distinct advantage in its enhanced capacity to capture long-range correlations compared to conventional recurrent neural networks (RNN). This characteristic makes it particularly well-suited for the precise identification of emotions from samples of handwriting and drawings, signifying a notable advancement in the field of emotion detection. The proposed method produced cutting-edge outcomes of 92.64% on the benchmark dataset known as EMOTHAW (Emotion Recognition via Handwriting and Drawing).
Collapse
|
6
|
From slides to insights: Harnessing deep learning for prognostic survival prediction in human colorectal cancer histology. Open Life Sci 2023; 18:20220777. [PMID: 38152577 PMCID: PMC10751997 DOI: 10.1515/biol-2022-0777] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/16/2023] [Revised: 10/24/2023] [Accepted: 10/26/2023] [Indexed: 12/29/2023] Open
Abstract
Prognostic survival prediction in colorectal cancer (CRC) plays a crucial role in guiding treatment decisions and improving patient outcomes. In this research, we explore the application of deep learning techniques to predict survival outcomes based on histopathological images of human colorectal cancer. We present a retrospective multicenter study utilizing a dataset of 100,000 nonoverlapping image patches from hematoxylin & eosin-stained histological images of CRC and normal tissue. The dataset includes diverse tissue classes such as adipose, background, debris, lymphocytes, mucus, smooth muscle, normal colon mucosa, cancer-associated stroma, and colorectal adenocarcinoma epithelium. To perform survival prediction, we employ various deep learning architectures, including convolutional neural network, DenseNet201, InceptionResNetV2, VGG16, VGG19, and Xception. These architectures are trained on the dataset using a multicenter retrospective analysis approach. Extensive preprocessing steps are undertaken, including image normalization using Macenko's method and data augmentation techniques, to optimize model performance. The experimental findings reveal promising results, demonstrating the effectiveness of deep learning models in prognostic survival prediction. Our models achieve high accuracy, precision, recall, and validation metrics, showcasing their ability to capture relevant histological patterns associated with prognosis. Visualization techniques are employed to interpret the models' decision-making process, highlighting important features and regions contributing to survival predictions. The implications of this research are manifold. The accurate prediction of survival outcomes in CRC can aid in personalized medicine and clinical decision-making, facilitating tailored treatment plans for individual patients. The identification of important histological features and biomarkers provides valuable insights into disease mechanisms and may lead to the discovery of novel prognostic indicators. The transparency and explainability of the models enhance trust and acceptance, fostering their integration into clinical practice. Research demonstrates the potential of deep learning models for prognostic survival prediction in human colorectal cancer histology. The findings contribute to the understanding of disease progression and offer practical applications in personalized medicine. By harnessing the power of deep learning and histopathological analysis, we pave the way for improved patient care, clinical decision support, and advancements in prognostic prediction in CRC.
Collapse
|
7
|
Optimised feature selection-driven convolutional neural network using gray level co-occurrence matrix for detection of cervical cancer. Open Life Sci 2023; 18:20220770. [PMID: 38045489 PMCID: PMC10693012 DOI: 10.1515/biol-2022-0770] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/05/2023] [Revised: 10/09/2023] [Accepted: 10/23/2023] [Indexed: 12/05/2023] Open
Abstract
Cervical cancer is one of the most dangerous and widespread illnesses afflicting women throughout the globe, particularly in East Africa and South Asia. In industrialised nations, the incidence of cervical cancer has consistently decreased over the past few decades. However, in developing countries, the reduction in incidence has been considerably slower, and in some instances, the incidence has increased. Implementing routine screenings for cervical cancer is something that has to be done to protect the health of women. Cervical cancer is famously difficult to diagnose and cure due to the slow rate at which it spreads and develops into more advanced stages of the disease. Screening for cervical cancer using a Pap smear, more often referred to as a Pap test, has the potential to detect the illness in its earlier stages. For the purpose of selecting features for this article, a gray level co-occurrence matrix (GLCM) technique was used. Following this step, classification is performed with methods such as convolutional neural network (CNN), support vector machine, and auto encoder. According to the findings of this experiment, the GLCM-CNN classifier proved to be the one with the highest degree of precision.
Collapse
|
8
|
A classifier model for prostate cancer diagnosis using CNNs and transfer learning with multi-parametric MRI. Front Oncol 2023; 13:1225490. [PMID: 38023149 PMCID: PMC10666634 DOI: 10.3389/fonc.2023.1225490] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/19/2023] [Accepted: 10/16/2023] [Indexed: 12/01/2023] Open
Abstract
Prostate cancer (PCa) is a major global concern, particularly for men, emphasizing the urgency of early detection to reduce mortality. As the second leading cause of cancer-related male deaths worldwide, precise and efficient diagnostic methods are crucial. Due to high and multiresolution MRI in PCa, computer-aided diagnostic (CAD) methods have emerged to assist radiologists in identifying anomalies. However, the rapid advancement of medical technology has led to the adoption of deep learning methods. These techniques enhance diagnostic efficiency, reduce observer variability, and consistently outperform traditional approaches. Resource constraints that can distinguish whether a cancer is aggressive or not is a significant problem in PCa treatment. This study aims to identify PCa using MRI images by combining deep learning and transfer learning (TL). Researchers have explored numerous CNN-based Deep Learning methods for classifying MRI images related to PCa. In this study, we have developed an approach for the classification of PCa using transfer learning on a limited number of images to achieve high performance and help radiologists instantly identify PCa. The proposed methodology adopts the EfficientNet architecture, pre-trained on the ImageNet dataset, and incorporates three branches for feature extraction from different MRI sequences. The extracted features are then combined, significantly enhancing the model's ability to distinguish MRI images accurately. Our model demonstrated remarkable results in classifying prostate cancer, achieving an accuracy rate of 88.89%. Furthermore, comparative results indicate that our approach achieve higher accuracy than both traditional hand-crafted feature techniques and existing deep learning techniques in PCa classification. The proposed methodology can learn more distinctive features in prostate images and correctly identify cancer.
Collapse
|
9
|
AD-CAM: Enhancing Interpretability of Convolutional Neural Networks with a Lightweight Framework - From Black Box to Glass Box. IEEE J Biomed Health Inform 2023; PP:1-14. [PMID: 37910403 DOI: 10.1109/jbhi.2023.3329231] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/03/2023]
Abstract
In the realm of machine vision, the convolutional neural network (CNN) is a frequently used and significant deep learning method. It is challenging to comprehend how predictions are formed since the inner workings of CNNs are sometimes seen as a black box. As a result, there has been an increase in interest among AI experts in creating AI systems that are easier to understand. Many strategies have shown promise in improving the interpretability of CNNs, including Class Activation Map (CAM), Grad-CAM, LIME, and other CAM-based approaches. These methods do, however, have certain drawbacks, such as architectural constraints or the requirement for gradient computations. We provide a simple framework termed Adaptive Learning based CAM (Adaptive-CAM) to take advantage of the connection between activation maps and network predictions. This framework includes temporarily masking particular feature maps. According to the Average Drop-Coherence-Complexity (ADCC) metrics, our method outperformed Score-CAM and another CAM-based activation map strategy in Residual Network-based models. With the exception of the VGG16 model, which witnessed a 1.94% decline in performance, the performance improvement spans from 3.78% to 7.72%. Additionally, Adaptive-CAM generates saliency maps that are on par with CAM-based methods and around 153 times superior to other CAM-based methods.
Collapse
|
10
|
Internet of Things-based sustainable environment management for large indoor facilities. PeerJ Comput Sci 2023; 9:e1623. [PMID: 37869451 PMCID: PMC10588707 DOI: 10.7717/peerj-cs.1623] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/10/2023] [Accepted: 09/08/2023] [Indexed: 10/24/2023]
Abstract
Due to global warming and climate change, the poultry industry is heavily impacted, especially the broiler industry, due to the sensitive immune system of broiler chickens. However, the continuous monitoring and controlling of the farm's environmental parameters can help to curtail the negative impacts of the environment on chickens' health, leading to increased meat production. This article presents smart solutions to such issues, which are practically implemented, and have low production and operational costs. In this article, an Internet of Things (IoT) based environmental parameters monitoring has been demonstrated for the poultry farmhouse. This system enables the collection and visualization of crucially sensed data automatically and reliably, and at a low cost to efficiently manage and operate a poultry farm. The proposed IoT-based remote monitoring system collects and visualizes environmental parameters, such as air temperature, relative humidity (RH), oxygen level (O2), carbon dioxide (CO2), carbon monoxide (CO), and ammonia (NH3) gas concentrations. The wireless sensor nodes have been designed and deployed for efficient data collection of the essential environmental parameters that are key for monitoring and decision-making process. The hardware is implemented and deployed successfully at a site within the control shed of the poultry farmhouse. The results revealed important findings related to the environmental conditions within the poultry farm. The temperature inside the control sheds remained within the desired range throughout the monitoring period, with daily average values ranging from 32 °C to 34 °C. The RH showed slight variations monitoring period, ranging from 65% to 75%, with a daily average of 70%. The O2 concentration exhibited an average value of 17% to 18.5% throughout the monitoring period. The CO2 levels showed occasional increases, reaching a maximum value of 1,100 ppm. However, this value was below the maximum permissible level of 2,500 ppm, indicating that the ventilation system was effective in maintaining acceptable CO2 levels within the control sheds. The NH3 gas concentration remained consistently low throughout the duration, with an average value of 50 parts per million (ppm).
Collapse
|
11
|
Multi-horizon short-term load forecasting using hybrid of LSTM and modified split convolution. PeerJ Comput Sci 2023; 9:e1487. [PMID: 37810340 PMCID: PMC10557505 DOI: 10.7717/peerj-cs.1487] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/06/2023] [Accepted: 06/16/2023] [Indexed: 10/10/2023]
Abstract
Precise short-term load forecasting (STLF) plays a crucial role in the smooth operation of power systems, future capacity planning, unit commitment, and demand response. However, due to its non-stationary and its dependency on multiple cyclic and non-cyclic calendric features and non-linear highly correlated metrological features, an accurate load forecasting with already existing techniques is challenging. To overcome this challenge, a novel hybrid technique based on long short-term memory (LSTM) and a modified split-convolution (SC) neural network (LSTM-SC) is proposed for single-step and multi-step STLF. The concatenating order of LSTM and SC in the proposed hybrid network provides an excellent capability of extraction of sequence-dependent features and other hierarchical spatial features. The model is evaluated by the Pakistan National Grid load dataset recorded by the National Transmission and Dispatch Company (NTDC). The load data is pre-processed and multiple other correlated features are incorporated into the data for performance enhancement. For generalization capability, the performance of LSTM-SC is evaluated on publicly available datasets of American Electric Power (AEP) and Independent System Operator New England (ISO-NE). The effect of temperature, a highly correlated input feature, on load forecasting is investigated either by removing the temperature or adding a Gaussian random noise into it. The performance evaluation in terms of RMSE, MAE, and MAPE of the proposed model on the NTDC dataset are 500.98, 372.62, and 3.72% for multi-step while 322.90, 244.22, and 2.38% for single-step load forecasting. The result shows that the proposed method has less forecasting error, strong generalization capability, and satisfactory performance on multi-horizon.
Collapse
|
12
|
Deep residual-dense network based on bidirectional recurrent neural network for atrial fibrillation detection. Sci Rep 2023; 13:15109. [PMID: 37704659 PMCID: PMC10499947 DOI: 10.1038/s41598-023-40343-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/23/2023] [Accepted: 08/09/2023] [Indexed: 09/15/2023] Open
Abstract
Atrial fibrillation easily leads to stroke, cerebral infarction and other complications, which will seriously harm the life and health of patients. Traditional deep learning methods have weak anti-interference and generalization ability. Therefore, we propose a new-fashioned deep residual-dense network via bidirectional recurrent neural network (RNN) model for atrial fibrillation detection. The combination of one-dimensional dense residual network and bidirectional RNN for atrial fibrillation detection simplifies the tedious feature extraction steps, and constructs the end-to-end neural network to achieve atrial fibrillation detection through data feature learning. Meanwhile, the attention mechanism is utilized to fuse the different features and extract the high-value information. The accuracy of the experimental results is 97.72%, the sensitivity and specificity are 93.09% and 98.71%, respectively compared with other methods.
Collapse
|
13
|
Hybrid Diagnostic Model for Improved COVID-19 Detection in Lung Radiographs Using Deep and Traditional Features. Biomimetics (Basel) 2023; 8:406. [PMID: 37754157 PMCID: PMC10526442 DOI: 10.3390/biomimetics8050406] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/22/2023] [Revised: 08/25/2023] [Accepted: 08/29/2023] [Indexed: 09/28/2023] Open
Abstract
A recently discovered coronavirus (COVID-19) poses a major danger to human life and health across the planet. The most important step in managing and combating COVID-19 is to accurately screen and diagnose affected people. The imaging technology of lung X-ray is a useful imaging identification/detection approach among them. The help of such computer-aided machines and diagnoses to examine lung X-ray images of COVID-19 instances can give supplemental assessment ideas to specialists, easing their workload to some level. The novel concept of this study is a hybridized approach merging pertinent manual features with deep spatial features for the classification of COVID-19. Further, we employed traditional transfer learning techniques in this investigation, utilizing four different pre-trained CNN-based deep learning models, with the Inception model showing a reasonably accurate result and a diagnosis accuracy of 82.17%. We provide a successful diagnostic approach that blends deep characteristics with machine learning classification to further increase clinical performance. It employs a complete diagnostic model. Two datasets were used to test the suggested approach, and it did quite well on several of them. On 1102 lung X-ray scans, the model was originally evaluated. The results of the experiments indicate that the suggested SVM model has a diagnostic accuracy of 95.57%. When compared to the Xception model's baseline, the diagnostic accuracy had risen by 17.58 percent. The sensitivity, specificity, and AUC of the proposed models were 95.37 percent, 95.39%, and 95.77%, respectively. To show the adaptability of our approach, we also verified our proposed model on other datasets. Finally, we arrived at results that were conclusive. When compared to research of a comparable kind, our suggested CNN model has a greater accuracy of classification and diagnostic effectiveness.
Collapse
|
14
|
A Novel Heteromorphous Convolutional Neural Network for Automated Assessment of Tumors in Colon and Lung Histopathology Images. Biomimetics (Basel) 2023; 8:370. [PMID: 37622975 PMCID: PMC10452605 DOI: 10.3390/biomimetics8040370] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/10/2023] [Revised: 07/31/2023] [Accepted: 08/03/2023] [Indexed: 08/26/2023] Open
Abstract
The automated assessment of tumors in medical image analysis encounters challenges due to the resemblance of colon and lung tumors to non-mitotic nuclei and their heteromorphic characteristics. An accurate assessment of tumor nuclei presence is crucial for determining tumor aggressiveness and grading. This paper proposes a new method called ColonNet, a heteromorphous convolutional neural network (CNN) with a feature grafting methodology categorically configured for analyzing mitotic nuclei in colon and lung histopathology images. The ColonNet model consists of two stages: first, identifying potential mitotic patches within the histopathological imaging areas, and second, categorizing these patches into squamous cell carcinomas, adenocarcinomas (lung), benign (lung), benign (colon), and adenocarcinomas (colon) based on the model's guidelines. We develop and employ our deep CNNs, each capturing distinct structural, textural, and morphological properties of tumor nuclei, to construct the heteromorphous deep CNN. The execution of the proposed ColonNet model is analyzed by its comparison with state-of-the-art CNNs. The results demonstrate that our model surpasses others on the test set, achieving an impressive F1 score of 0.96, sensitivity and specificity of 0.95, and an area under the accuracy curve of 0.95. These outcomes underscore our hybrid model's superior performance, excellent generalization, and accuracy, highlighting its potential as a valuable tool to support pathologists in diagnostic activities.
Collapse
|
15
|
A residual connection enabled deep neural network model for optic disk and optic cup segmentation for glaucoma diagnosis. Sci Prog 2023; 106:368504231201329. [PMID: 37743660 PMCID: PMC10521305 DOI: 10.1177/00368504231201329] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 09/26/2023]
Abstract
Glaucoma diagnosis at an early stage is vital for the timely initiation of its treatment for and preventing possible vision loss. For glaucoma diagnosis, an accurate estimation of the cup-to-disk ratio (CDR) is required. The current automatic CDR computation techniques attribute lower accuracy and higher complexity, which are important considerations for diagnostics system design to be used for such critical diagnoses. The current methods involve a deeper deep learning model, comprising a large number of parameters, which results in higher system complexity and training/testing time. To address these challenges, this paper proposes a Residual Connection (non-identity)-based Deep Neural Network (RC-DNN), which is based on non-identity residual connectivity for joint optic disk (OD) and optic cup (OC) detection. The proposed model is emboldened by efficient residual connectivity, which is beneficial in several ways. First, the model is efficient and can perform simultaneous segmentation of the OC and OD. Second, the efficient residual information flow permeates the vanishing gradient problem which results in faster converges of the model. Third, feature inspiration empowers the network to perform the segmentation with only a few network layers. We performed a comprehensive performance evaluation of the developed model based on its training in RIM-ONE and DRISHTIGS databases. For OC segmentation, for the images (test set) from {DRISHTI-GS and RIM-ONE} datasets, our proposed model achieves the dice coefficient, Jaccard coefficient, sensitivity, specificity, and balanced accuracy of {92.62, 86.52}, {86.87, 77.54}, {94.21, 95.36}, {99.83, 99.639}, and {94.2, 98.9}, respectively. These experimental results indicate that the developed model provides significant performance enhancement for joint OC and OD segmentation. Additionally, the reduced computational complexity based on reduced model parameters and higher segmentation accuracy provides the additional features of efficacy, robustness, and reliability of the developed model. These attributes of the developed model advocate for its deployment of population-scale glaucoma screening programs.
Collapse
|
16
|
Abstract
AbstractMulticlass classification of brain tumors is an important area of research in the field of medical imaging. Since accuracy is crucial in the classification, a number of techniques are introduced by computer vision researchers; however, they still face the issue of low accuracy. In this article, a new automated deep learning method is proposed for the classification of multiclass brain tumors. To realize the proposed method, the Densenet201 Pre-Trained Deep Learning Model is fine-tuned and later trained using a deep transfer of imbalanced data learning. The features of the trained model are extracted from the average pool layer, which represents the very deep information of each type of tumor. However, the characteristics of this layer are not sufficient for a precise classification; therefore, two techniques for the selection of features are proposed. The first technique is Entropy–Kurtosis-based High Feature Values (EKbHFV) and the second technique is a modified genetic algorithm (MGA) based on metaheuristics. The selected features of the GA are further refined by the proposed new threshold function. Finally, both EKbHFV and MGA-based features are fused using a non-redundant serial-based approach and classified using a multiclass SVM cubic classifier. For the experimental process, two datasets, including BRATS2018 and BRATS2019, are used without increase and have achieved an accuracy of more than 95%. The precise comparison of the proposed method with other neural nets shows the significance of this work.
Collapse
|
17
|
Human Behavior Analysis Based on Multi-Types Features Fusion and Von Nauman Entropy Based Features Reduction. JOURNAL OF MEDICAL IMAGING AND HEALTH INFORMATICS 2019. [DOI: 10.1166/jmihi.2019.2611] [Citation(s) in RCA: 26] [Impact Index Per Article: 5.2] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/23/2022]
|
18
|
Towards Void Hole Alleviation by Exploiting the Energy Efficient Path and by Providing the Interference-Free Proactive Routing Protocols in IoT Enabled Underwater WSNs. SENSORS 2019; 19:s19061313. [PMID: 30884749 PMCID: PMC6471835 DOI: 10.3390/s19061313] [Citation(s) in RCA: 12] [Impact Index Per Article: 2.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 02/25/2019] [Accepted: 02/25/2019] [Indexed: 11/16/2022]
Abstract
Nowadays, the Internet of Things enabled Underwater Wireless Sensor Network (IoT-UWSN) is suffering from serious performance restrictions, i.e., high End to End (E2E) delay, low energy efficiency, low data reliability, etc. The necessity of efficient, reliable, collision and interference-free communication has become a challenging task for the researchers. However, the minimum Energy Consumption (EC) and low E2E delay increase the performance of the IoT-UWSN. Therefore, in the current work, two proactive routing protocols are presented, namely: Bellman⁻Ford Shortest Path-based Routing (BF-SPR-Three) and Energy-efficient Path-based Void hole and Interference-free Routing (EP-VIR-Three). Then we formalized the aforementioned problems to accomplish the reliable data transmission in Underwater Wireless Sensor Network (UWSN). The main objectives of this paper include minimum EC, interference-free transmission, void hole avoidance and high Packet Delivery Ratio (PDR). Furthermore, the algorithms for the proposed routing protocols are presented. Feasible regions using linear programming are also computed for optimal EC and to enhance the network lifespan. Comparative analysis is also performed with state-of-the-art proactive routing protocols. In the end, extensive simulations have been performed to authenticate the performance of the proposed routing protocols. Results and discussion disclose that the proposed routing protocols outperformed the counterparts significantly.
Collapse
|
19
|
CoSiM-RPO: Cooperative Routing with Sink Mobility for Reliable and Persistent Operation in Underwater Acoustic Wireless Sensor Networks. SENSORS 2019; 19:s19051101. [PMID: 30836710 PMCID: PMC6427485 DOI: 10.3390/s19051101] [Citation(s) in RCA: 10] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 01/14/2019] [Revised: 02/18/2019] [Accepted: 02/22/2019] [Indexed: 11/16/2022]
Abstract
An efficient algorithm for the persistence operation of data routing is crucial due to the uniqueness and challenges of the aqueous medium of the underwater acoustic wireless sensor networks (UA-WSNs). The existing multi-hop algorithms have a high energy cost, data loss, and less stability due to many forwarders for a single-packet delivery. In order to tackle these constraints and limitations, two algorithms using sink mobility and cooperative technique for UA-WSNs are devised. The first one is sink mobility for reliable and persistence operation (SiM-RPO) in UA-WSNs, and the second is the enhanced version of the SiM-RPO named CoSiM-RPO, which utilizes the cooperative technique for better exchanging of the information and minimizes data loss probability. To cover all of the network through mobile sinks (MSs), the division of the network into small portions is accomplished. The path pattern is determined for MSs in a manner to receive data even from a single node in the network. The MSs pick the data directly from the nodes and check them for the errors. When erroneous data are received at the MS, then the relay cooperates to receive correct data. The proposed algorithm boosts the network lifespan, throughput, delay, and stability more than the existing counterpart schemes.
Collapse
|
20
|
An implementation of normal distribution based segmentation and entropy controlled features selection for skin lesion detection and classification. BMC Cancer 2018; 18:638. [PMID: 29871593 PMCID: PMC5989438 DOI: 10.1186/s12885-018-4465-8] [Citation(s) in RCA: 67] [Impact Index Per Article: 11.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/05/2017] [Accepted: 04/30/2018] [Indexed: 01/28/2023] Open
Abstract
Background Melanoma is the deadliest type of skin cancer with highest mortality rate. However, the annihilation in its early stage implies a high survival rate therefore, it demands early diagnosis. The accustomed diagnosis methods are costly and cumbersome due to the involvement of experienced experts as well as the requirements for the highly equipped environment. The recent advancements in computerized solutions for this diagnosis are highly promising with improved accuracy and efficiency. Methods In this article, a method for the identification and classification of the lesion based on probabilistic distribution and best features selection is proposed. The probabilistic distribution such as normal distribution and uniform distribution are implemented for segmentation of lesion in the dermoscopic images. Then multi-level features are extracted and parallel strategy is performed for fusion. A novel entropy-based method with the combination of Bhattacharyya distance and variance are calculated for the selection of best features. Only selected features are classified using multi-class support vector machine, which is selected as a base classifier. Results The proposed method is validated on three publicly available datasets such as PH2, ISIC (i.e. ISIC MSK-2 and ISIC UDA), and Combined (ISBI 2016 and ISBI 2017), including multi-resolution RGB images and achieved accuracy of 97.5%, 97.75%, and 93.2%, respectively. Conclusion The base classifier performs significantly better on proposed features fusion and selection method as compared to other methods in terms of sensitivity, specificity, and accuracy. Furthermore, the presented method achieved satisfactory segmentation results on selected datasets.
Collapse
|
21
|
Retransmission Avoidance for Reliable Data Delivery in Underwater WSNs. SENSORS 2018; 18:s18010149. [PMID: 29316664 PMCID: PMC5795748 DOI: 10.3390/s18010149] [Citation(s) in RCA: 14] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 11/14/2017] [Revised: 12/27/2017] [Accepted: 01/04/2018] [Indexed: 12/01/2022]
Abstract
The energy-efficient and reliable delivery of data packets in resource constraint underwater wireless sensor networks (UWSNs) is one of the key considerations to enhance the network lifetime. The traditional re-transmissions approach consumes the node battery and increases the communication overhead, which results in congestion and affects the reliable data packet delivery in the network. To ensure the reliability and conserve the node battery, in this paper, we propose adaptive forwarding layer multipath power control routing protocol to reduce the energy dissipation, achieve the data reliability and avoid the energy hole problem. In order to achieve the reliability, tree based topology is exploited to direct multiple copies of the data packet towards the surface through cross nodes in the network. The energy dissipation is reduced by a substantial amount with the selection of low noise path between the source and the destination including the information of neighbors of the potential forwarder node. Extensive simulation results show that our proposed work outperforms the compared existing scheme in terms of energy efficiency and packet received ratio (PRR).
Collapse
|