Search

Owing to the tiny abnormal developments in breast masses, examination and detection of breast cancer utilizing high-resolution images havebecome a complex process. Thus, an innovative framework for breast cancer classification is constructed, named SqueezeNet-Fractional Pelican African Vulture Optimization (FPAVO). Here, preprocessingis done utilizing a median filter and Region of Interest (ROI) extrication. The segmentation process is done using a hybrid network classifier, Recurrent Prototypical Network (RP-Net) and KNet,which is trained by the Pelican African Vulture Optimization (PAVO). The PAVO is the consolidation of the Pelican Optimization Algorithm (POA) and African Vulture Optimization (AVO). Then,texture and statistical features are extricated. After that, the data augmentation is done. Then, the SqueezeNet is used for classification, and it is trained by FPAVO, which integrates the fractionalcalculus (FC) into PAVO. The SqueezeNet-FPAVO produced high testing accuracy, sensitivity, specificity and F-measure of 0.920, 0.924, 0.925, and 0.897. [ABSTRACT FROM AUTHOR]

The composition of web service is the best chance provided by Service-Oriented Computing and Service-Oriented Architecture as it gives real competitive benefits for some industrial and technological actors via presenting them with the probability to guarantee fast and inexpensive improvement of collaborative and distributed software applications. Here, a novel technique is introduced, which contains several phases for better web services. At first, the requirements specification phase is enabled with a set of requirements such as non-functional and functional requirements. Next to the requirements specification stage, the discovery stage is enabled to choose the suitable web services that have high-matching profiles with the developer's requirement set. Here, for a semantic matching algorithm, a new hybrid similarity measure is developed. Additionally, among the group of candidate services that the discovery phase returned, the best service is selected during the selection step. Then, hybrid Squeeze_Long Short-Term Memory (Squeeze_LSTM) is used for choosing the best service and it is designed by the formation of SqueezeNet and LSTM. The Semantic Web Services are finally implemented. The efficiency of the Squeeze_LSTM is evaluated and has achieved a superior precision of 0.909, recall of 0.890, and response time of 6.461S. [ABSTRACT FROM AUTHOR]

Rice (Oryza sativa) is a linchpin of global food security, feeding most of the world's population, especially in Asia and Africa. Nitrogen (N), Phosphorus (P), and Potassium (K) must be held in precise equilibrium for rice to grow right and produce well. A lack of these nutrients can put plant health at risk, causing poor yields that result in huge economic losses to farmers. This means it is important to have an accurate diagnosis done on time so that corrective measures can be taken to ensure the sustainability of rice cultivation. Conventional techniques used to detect nutrient deficiencies in rice plants like manual inspection and biochemical testing are frequently inadequate. Manual inspection may be effective, but it is laborious, subjective, and unworkable when working on a large-scale farm. On the other hand, the accuracy of biochemical tests does not compensate for their time-consuming nature, their cost as well as the need for specialized equipment and expertise; this makes them inaccessible to many smallholder farmers. This research proposes Attention SqueezeNet-a streamlined deep learning model aimed at overcoming such limitations. Attention SqueezeNet leverages the Kaggle "Nutrient Deficiency Symptoms in Rice" dataset by including an attention mechanism that focuses on significant visual features associated with N P K deficiency on rice leaves. Compared with traditional methods, therefore, these focused areas enable more accurate diagnostic results from the model as opposed to conventional methods which diagnose diseases broadly rather than focusing on specific symptoms. In this context, we propose: i) developing a robust Attention SqueezeNet model specifically aimed at categorizing nutrient deficiencies in rice plants; ii) making the model more robust against unseen data variations through pre-processing and augmenting existing dataset; iii) comparing the performance of Attention SqueezeNet with existing deep learning models in terms of accuracy and efficiency per computational unit utilized (CU). The results show that Attention SqueezeNet has better classification accuracy compared to state-of-the-art models and is more computationally efficient than them. This has resulted in the transformation of agriculture through automation and objectivity in diagnosis of nutrient deficiency which ensures sustainable crop management practices towards global food security. [ABSTRACT FROM AUTHOR]

In recent years, the COVID-19 outbreak has affected humanity across the globe. The frequent symptoms of COVID-19 are identical to the normal flu, such as fever and cough. COVID-19 disseminates rapidly, and it has become a prominent cause of mortality. Nowadays, the new wave of COVID-19 has created significant impacts in China. This virus can have detrimental effects on people of all ages, particularly the elderly, due to their weak immune systems. The real-time polymerase chain reaction (RT-PCR) examination is typically performed for the identification of coronavirus. RT-PCR is an expensive and time requiring method, accompanied by a significant rate of false negative detections. Therefore, it is mandatory to develop an inexpensive, fast, and reliable method to detect COVID-19. X-ray images are generally utilized to detect diverse respiratory conditions like pulmonary infections, breathlessness syndrome, lung cancer, air collection in spaces of the lungs, etc. This study has also utilized a chest X-ray dataset to identify COVID-19 and pneumonia. In this research work, we proposed a novel deep learning model CP_DeepNet, which is based on a pre-trained deep learning model such as SqueezeNet, and further added three blocks of convolutional layers to it for assessing the classification efficacy. Furthermore, we employed a data augmentation method for generating more images to overcome the problem of model overfitting. We utilized COVID-19 radiograph dataset for evaluating the performance of the proposed model. To elaborate further, we obtained significant results with accuracy of 99.32%, a precision of 100%, a recall of 99%, a specificity of 99.2%, an area under the curve of 99.78%, and an F1-score of 99.49% on CP_DeepNet for the binary classification of COVID-19 and normal class. We also employed CP_DeepNet for the multiclass classification of COVID-19, pneumonia, and normal person, in which CP_DeepNet achieved accuracy, precision, recall, specificity, area under curve, and F1-score of 99.62%, 99.79%, 99.52%, 99.69, 99.62, and 99.72%, respectively. Comparative analysis of experimental results with different preexisting techniques shows that the proposed model is more dependable as compared to RT-PCR and other prevailing modern techniques for the detection of COVID-19. [ABSTRACT FROM AUTHOR]

A complicated neuro-developmental disorder called Autism Spectrum Disorder (ASD) is abnormal activities related to brain development. ASD generally affects the physical impression of the face as well as the growth of the brain in children. An early and proper medical diagnosis is essential for ASD affected children to enhance their quality of life. However, the clinical detection of ASD is a difficult task and time-consuming, hence it is essential to design an ASD detection approach for precise diagnosis of ASD children. In this research, an algorithmic approach called Fractional Tasmanian Rat Swarm Optimization driven SqueezeNet (FTRSO-SqueezeNet) is designed for the detection of ASD. The median filter and Region of Interest (RoI) extraction are used to de-noise the input image initially and extract a particular region from the filtered image. Later, the nub regions are extracted by choosing the optimal grid from the pre-processed image and using different feature extractors the features from the input images are determined. Finally, the detection of ASD is done by using SqueezeNet, which is trained using the FTRSO approach. The performance of the method is estimated, where the designed model achieved higher performance with an accuracy of 94.55%, sensitivity of 92.53%, and specificity of 95.22% than other prevailing approaches. [ABSTRACT FROM AUTHOR]

Abstract Human-Computer Interaction (HCI) is a multidisciplinary field focused on designing and utilizing computer technology, underlining the interaction interface between computers and humans. HCI aims to generate systems that allow consumers to relate to computers effectively, efficiently, and pleasantly. Multiple Spoken Language Identification (SLI) for HCI (MSLI for HCI) denotes the ability of a computer system to recognize and distinguish various spoken languages to enable more complete and handy interactions among consumers and technology. SLI utilizing deep learning (DL) involves using artificial neural networks (ANNs), a subset of DL models, to automatically detect and recognize the language spoken in an audio signal. DL techniques, particularly neural networks (NNs), have succeeded in various pattern detection tasks, including speech and language processing. This paper develops a novel Coot Optimizer Algorithm with a DL-Driven Multiple SLI and Detection (COADL-MSLID) technique for HCI applications. The COADL-MSLID approach aims to detect multiple spoken languages from the input audio regardless of gender, speaking style, and age. In the COADL-MSLID technique, the audio files are transformed into spectrogram images as a primary step. Besides, the COADL-MSLID technique employs the SqueezeNet model to produce feature vectors, and the COA is applied to the hyperparameter range of the SqueezeNet method. The COADL-MSLID technique exploits the SLID process’s convolutional autoencoder (CAE) model. To underline the importance of the COADL-MSLID technique, a series of experiments were conducted on the benchmark dataset. The experimentation validation of the COADL-MSLID technique exhibits a greater accuracy result of 98.33% over other techniques.

Abstract The question of whether artificial intelligence (AI) can surpass human capabilities is crucial in the application of AI in clinical medicine. To explore this, an interpretable deep learning (DL) model was developed to assess myopia status using retinal refraction maps obtained with a novel peripheral refractor. The DL model demonstrated promising performance, achieving an AUC of 0.9074 (95% CI 0.83–0.97), an accuracy of 0.8140 (95% CI 0.70–0.93), a sensitivity of 0.7500 (95% CI 0.51–0.90), and a specificity of 0.8519 (95% CI 0.68–0.94). Grad-CAM analysis provided interpretable visualization of the attention of DL model and revealed that the DL model utilized information from the central retina, similar to human readers. Additionally, the model considered information from vertical regions across the central retina, which human readers had overlooked. This finding suggests that AI can indeed surpass human capabilities, bolstering our confidence in the use of AI in clinical practice, especially in new scenarios where prior human knowledge is limited.

In the medical field, Internet of Things (IoT) applications allow for real-time diagnosis and remote patient monitoring, commonly called Internet of Health Things (IoHT). However, cybersecurity attacks may interrupt hospital operations and threaten patients’ health and well-being due to this integration. Hence, developing an Intrusion Detection System (IDS) suited explicitly for healthcare systems is essential to ensure efficiency and accuracy. Nevertheless, it is challenging to integrate anomaly-based IDS frameworks in healthcare systems as they necessitate additional processing time, temporal feature retention, and increased complexity. Therefore, a deep learning system based on SqueezeNet and NasNet is presented in this paper to detect intrusions in a healthcare setting. In this, SqueezeNet is employed to extract more significant features. On the other hand, network breaches while data transmission across distinct locations are detected by the NasNet-based classifier. In addition, the Rider Optimization Algorithm (ROA) is applied to adjust the classifier’s hyperparameters, guaranteeing that it would accurately detect attacks. Moreover, the Auxiliary Classifier Generative Adversarial Network (ACGAN) approach is integrated into the proposed framework to avoid data imbalance. Applying different performance constraints, the proposed approach is thoroughly assessed on three publicly available datasets (TON-IoT, ECU-IoHT, and WUSTL-EHMS). The results show that the proposed deep learning-based cybersecurity model outperforms traditional methods and produces better outcomes. [ABSTRACT FROM AUTHOR]

The detection of rice leaf disease is an essential step for implementing precise and timely interventions, thereby mitigating the spread and minimizing the ecological and economic consequences. This work proposes a deep learning-based biotic rice leaf disease detection model wherein ensemble models have been proposed using pre-trained models as feature extractors and machine learning/deep learning classifiers for performing classification. VGG16, SqueezeNet, and InceptionV3, have been used as feature extractors. The fine-tuning of various pre-trained models has been done by setting the hyperparameters to the prominent values for achieving optimal results. These hyperparameters include batch size, learning rate, optimizer, epochs, train, and test ratio. The model utilizes various machine learning and deep learning classifiers to perform the multiclass classification on extracted features from a pooled rice leaf disease dataset. In the proposed work, SqueezeNet with neural network classifier achieved the highest accuracy of 93.3%, Area Under Curve (AUC) of 0.989, precision of 0.928, recall of 0.931, and F1-score of 0.928. [ABSTRACT FROM AUTHOR]

Generally, melanoma skin disease is also a type of cancer, which is complex to determine. In case, various skin cancer diseases are identified at an early stage, then the death rate is to be reduced. Medical imaging technology acts as an important role in perceiving these types of skin lesions perfectly. This research introduced the automatic skin cancer detection technique, namely Aquila Driving Training-Based optimization (ADTBO) with Squeeze Net. The ADTBO is derived by the combination of Aquila Optimizer (AO) and Driving Training-Based optimization (DTBO). The skin cancer segmentation is completed by the Minkowski-based dual network, where the dual network encompasses the PsiNet and Deep Joint segmentation. The data augmentation and feature extraction are completed to enhance the effectiveness of detection. The ADTBO-SqueezeNet achieved the higher performance based on the accuracy of 0.976, specificity of 0.984 and sensitivity of 0.987. The segmentation accuracy achieved by the Minkowski-based Dual Network is 0.948. [ABSTRACT FROM AUTHOR]

Breast cancer with increased risk in women is identified with Breast Magnetic Resonance Imaging (Breast MRI) and this helps in evaluating treatment therapies. Breast MRI is time time-consuming process that involves the assessment of current imaging. This research work depends on the detection of breast cancer at the earlier stages. Among various cancers, breast cancer in women occurs in larger accounts for almost 30% of estimated cancer cases. In this research, many steps are followed for breast cancer detection like pre-processing, segmentation, augmentation, extraction of features, and cancer detection. Here, the median filter is utilized for pre-processing, as well as segmentation is followed after pre-processing, which is done by Psi-Net. Moreover, the process of augmentation like shearing, translation, and cropping are followed after segmentation. Also, the segmented image tends to process feature extraction, where features like shape features, Completed Local Binary Pattern (CLBP), Pyramid Histogram of Oriented Gradients (PHOG), and statistical features are extracted. Finally, breast cancer is detected using the DL model, SqueezeNet. Here, the newly devised Flamingo Search SailFish Optimizer (FSSFO) is used in training Psi-Net as well as SqueezeNet. Furthermore, FSSFO is the combination of both the Flamingo Search Algorithm (FSA) and SailFish Optimizer (SFO). [ABSTRACT FROM AUTHOR]

The majority of applications use automatic image recognition technologies to carry out a range of tasks. Therefore, it is crucial to identify and classify image distortions to improve image quality. Despite efforts in this area, there are still many challenges in accurately and reliably classifying distorted images. In this paper, we offer a comprehensive analysis of models of both non-lightweight and lightweight deep convolutional neural networks (CNNs) for the classification of distorted images. Subsequently, an effective method is proposed to enhance the overall performance of distortion image classification. This method involves selecting features from the pretrained models' capabilities and using a strong classifier. The experiments utilized the kadid10k dataset to assess the effectiveness of the results. The K-nearest neighbor (KNN) classifier showed better performance than the naïve classifier in terms of accuracy, precision, error rate, recall and F1 score. Additionally, SqueezeNet outperformed other deep CNN models, both lightweight and non-lightweight, across every evaluation metric. The experimental results demonstrate that combining SqueezeNet with KNN can effectively and accurately classify distorted images into the correct categories. The proposed SqueezeNet-KNN method achieved an accuracy rate of 89%. As detailed in the results section, the proposed method outperforms state-of-the-art methods in accuracy, precision, error, recall, and F1 score measures. [ABSTRACT FROM AUTHOR]

The question of whether artificial intelligence (AI) can surpass human capabilities is crucial in the application of AI in clinical medicine. To explore this, an interpretable deep learning (DL) model was developed to assess myopia status using retinal refraction maps obtained with a novel peripheral refractor. The DL model demonstrated promising performance, achieving an AUC of 0.9074 (95% CI 0.83–0.97), an accuracy of 0.8140 (95% CI 0.70–0.93), a sensitivity of 0.7500 (95% CI 0.51–0.90), and a specificity of 0.8519 (95% CI 0.68–0.94). Grad-CAM analysis provided interpretable visualization of the attention of DL model and revealed that the DL model utilized information from the central retina, similar to human readers. Additionally, the model considered information from vertical regions across the central retina, which human readers had overlooked. This finding suggests that AI can indeed surpass human capabilities, bolstering our confidence in the use of AI in clinical practice, especially in new scenarios where prior human knowledge is limited. Highlights: • Developing a deep learning algorithm to assess myopia status using retinal refraction maps obtained from a novel peripheral refractor. • The deep learning algorithm accurately determined myopia status with high accuracy. • Interpretable Grad-CAM analysis indicated that the deep learning algorithm utilized information from the central retina as well as vertical regions. [ABSTRACT FROM AUTHOR]

Cruise machinists operate in dynamic and physically demanding environments where improper posture can lead to musculoskeletal disorders, adversely affecting their health and work efficiency. Current ergonomic assessments in such settings are often generic and not tailored to the unique challenges of maritime operations. This paper presents a novel application of artificial intelligence tools for real-time posture estimation specifically designed for cruise machinists. The primary aim is to enhance occupational health and safety by providing precise, real-time feedback on ergonomic practices. We developed a dataset by capturing video recordings of cruise machinists at work, which were processed to extract skeletal outlines using advanced computer vision techniques. This dataset was used to train deep neural networks, optimizing them for accuracy in diverse and constrained shipboard environments. The networks were tested across various computational platforms to ensure robustness and adaptability. The AI model demonstrated high efficacy in recognizing both correct and incorrect postures under real-world conditions aboard ships. The system significantly outperformed traditional ergonomic assessment tools in terms of speed, accuracy, and the ability to provide instant feedback. The findings suggest that AI-enhanced ergonomic assessments could be a transformative approach for occupational health across various industries. [ABSTRACT FROM AUTHOR]

Optical character recognition (OCR) is a process that recognizes and converts data from scanned images, including both handwritten and printed documents, into an accessible format. The challenges in Telugu OCR arise from compound characters, an extensive character set, limited datasets, character similarities, and difficulties in segmenting overlapping characters. To tackle these segmentation complexities, an algorithm has been developed, prioritizing the preservation of essential features during character segmentation. For distinguishing between structurally similar characters, we used convolutional neural networks (CNN) due to their feature-extracting properties. We have employed the CNN model, the SqueezeNet for feature extraction, resulting in an impressive character recognition rate of 94% and a word recognition rate of 80%. [ABSTRACT FROM AUTHOR]

Operating in harsh environments, drilling pumps are highly susceptible to failure and challenging to diagnose. To enhance the fault diagnosis accuracy of the drilling pump fluid end and ensure the safety and stability of drilling operations, this paper proposes a fault diagnosis method based on time-frequency analysis and convolutional neural networks. Firstly, continuous wavelet transform (CWT) is used to convert the collected vibration signals into time-frequency diagrams, providing a comprehensive database for fault diagnosis. Next, a SqueezeNet-based fault diagnosis model is developed to identify faults. To validate the effectiveness of the proposed method, fault signals from the fluid end were collected, and fault diagnosis experiments were conducted. The experimental results demonstrated that the proposed method achieved an accuracy of 97.77% in diagnosing nine types of faults at the fluid end, effectively enabling precise fault diagnosis, which is higher than the accuracy of a 1D convolutional neural network by 14.55%. This study offers valuable insights into the fault diagnosis of drilling pumps and other complex equipment. [ABSTRACT FROM AUTHOR]

Abstract The maize leaf diseases create severe yield reductions and critical problems. The maize leaf disease should be discovered early, perfectly identified, and precisely diagnosed to make greater yield. This work studies three main leaf diseases: common rust, blight, and grey leaf spot. This approach involves pre-processing, including sampling and labelling, while ensuring class balance and preventing overfitting via the SMOTE algorithm. The maize leaf dataset with augmentation was used to classify these diseases using several deep-learning pre-trained networks, including VGG16, Resnet34, Resnet50, and SqueezeNet. The model was evaluated using a maize leaf dataset that included various leaf classes, mini-batch sizes, and input sizes. Performance measures, recall, precision, accuracy, F1-score, and confusion matrix were computed for each network. The SqueezeNet learning model produces an accuracy of 97% in classifying four different classes of plant leaf datasets. Comparatively, the SqueezeNet learning model has improved accuracy by 2–5% and reduced the mean square error by 4–11% over VGG16, Resnet34, and Resnet50 deep learning models.

Facial emotion recognition (FER) is largely utilized to analyze human emotion in order to address the needs of many real-time applications such as computer-human interfaces, emotion detection, forensics, biometrics, and human-robot collaboration. Nonetheless, existing methods are mostly unable to offer correct predictions with a minimum error rate. In this paper, an innovative facial emotion recognition framework, termed extended walrus-based deep learning with Botox feature selection network (EWDL-BFSN), was designed to accurately detect facial emotions. The main goals of the EWDL-BFSN are to identify facial emotions automatically and effectively by choosing the optimal features and adjusting the hyperparameters of the classifier. The gradient wavelet anisotropic filter (GWAF) can be used for image pre-processing in the EWDL-BFSN model. Additionally, SqueezeNet is used to extract significant features. The improved Botox optimization algorithm (IBoA) is then used to choose the best features. Lastly, FER and classification are accomplished through the use of an enhanced optimization-based kernel residual 50 (EK-ResNet50) network. Meanwhile, a nature-inspired metaheuristic, walrus optimization algorithm (WOA) is utilized to pick the hyperparameters of EK-ResNet50 network model. The EWDL-BFSN model was trained and tested with publicly available CK+ and FER-2013 datasets. The Python platform was applied for implementation, and various performance metrics such as accuracy, sensitivity, specificity, and F1-score were analyzed with state-of-the-art methods. The proposed EWDL-BFSN model acquired an overall accuracy of 99.37 and 99.25% for both CK+ and FER-2013 datasets and proved its superiority in predicting facial emotions over state-of-the-art methods.

Globally, eye disorders have been a major issue, especially in developing nations where resources for technology and financing are limited. Due to its tremendous feature learning ability, CNN has obtained remarkable progress in the area of fundus images. Through appropriate analysis and examination of fundus images, computer-aided diagnosis may yield information with a standard value for experts in clinical diagnosis or screening. However, the majority of earlier investigations have focused on the identification of a particular fundus disease, and the accurate fast classification of multiple fundus diseases remains a major challenge. Extremely numerous retinal fundus images must be analyzed to achieve a categorization that is reliable, fast, and precise. As a result, this research aims to propose a novel classification model for eye illnesses based on four major steps: (a) Pre-processing, (b) segmentation of blood vessels, (c) Extracting features and (d) Multi-classification of eye disease. In order to accurately classify the images, it is necessary to extract the essential informational characteristics from the segmented blood vessels. The categorization of eye diseases is carried out using a hybrid classifier that integrates the SqueezeNet and Long-Term Recurrent Convolutional Network (LRCN). We trained and validated our models using tenfold cross-validation tests on a database of fundus images including five categories: Normal, Retinitis Pigmentosa, Pathological Myopia, Maculopathy and Glaucoma. The efficacy of the presented approach is evaluated based on specificity, F1-score, recall, precision and accuracy of 98%, 97.8%, 97.6%, 98.4% and 98% respectively. [ABSTRACT FROM AUTHOR]

The maize leaf diseases create severe yield reductions and critical problems. The maize leaf disease should be discovered early, perfectly identified, and precisely diagnosed to make greater yield. This work studies three main leaf diseases: common rust, blight, and grey leaf spot. This approach involves pre-processing, including sampling and labelling, while ensuring class balance and preventing overfitting via the SMOTE algorithm. The maize leaf dataset with augmentation was used to classify these diseases using several deep-learning pre-trained networks, including VGG16, Resnet34, Resnet50, and SqueezeNet. The model was evaluated using a maize leaf dataset that included various leaf classes, mini-batch sizes, and input sizes. Performance measures, recall, precision, accuracy, F1-score, and confusion matrix were computed for each network. The SqueezeNet learning model produces an accuracy of 97% in classifying four different classes of plant leaf datasets. Comparatively, the SqueezeNet learning model has improved accuracy by 2–5% and reduced the mean square error by 4–11% over VGG16, Resnet34, and Resnet50 deep learning models. [ABSTRACT FROM AUTHOR]

Bearings are vital components in rotating machinery. Undetected bearing faults may result not only in financial loss, but also in the loss of lives. Hence, there exists an abundance of studies working on the early detection of bearing faults. The rising use of deep learning in recent years increased the number of imaging types/neural network architectures used for bearing fault classification, making it challenging to choose the most suitable 2-D imaging method and neural network. This study aims to address this challenge, by sharing the results of the training of eighteen imaging methods with four different networks using the same vibration data and training metrics. To further strengthen the results, the validation dataset size was taken as five times the training dataset size. The best results obtained is 99.89% accuracy by using Scattergram Filter Bank 1 as the image input, and ResNet-50 as the network for training. Prior to our work, Scattergram images have never been used for bearing fault classification. Ten out of 72 methods used in this work resulted in accuracies higher than 99.5%. [ABSTRACT FROM AUTHOR]

Soil color recognition through AI approaches is crucial for efficient and rapid soil analysis in geotechnical engineering. It enables improved site characterization and supports informed decision-making in engineering projects. The integration of soil color data with other geospatial information enhances geological mapping and modeling capabilities. The expansion of deep learning can assist different stockholders in soil classification as a significant study in recent years. Therefore, this paper presented an efficient IoT-based hybrid CNN-SVM model to classify five soil classes through soil images. The recommended framework utilizes a hybrid model based on CNN as feature selection. After that, a multiclass SVM as a classifier is utilized for the soil classification task in an effective manner as well as new real-time IoT-based portable soil detection devices for geotechnical engineers in Geo-sites. The proposed framework is estimated using a dataset that comprehends a total of 252 soil images for investigation purposes with different evaluation metrics. The proposed hybrid framework using CNN models such as Squeezenet, Alexnet, and Resnet50 with the multiclass SVM classifier gives good accuracy of 86%, 96%, and 95%, respectively. In contrast, the CNN models only give accuracy of 80%, 89%, and 87% for the Squeezenet, Alexnet, and Resnet50, respectively. From the obtained results, we observed that the offered hybrid Alexnet-SVM gives a higher performance while the Squeezenet-SVM gives the lowest concert. Overall, the attained results showed that the recommended hybrid framework gives the best concert for soil classification, which can help in making an efficient support decision-making system for real-time geotechnical engineering applications. [ABSTRACT FROM AUTHOR]

In this paper, a clear underwater image is attained by a fusion process using Transfer Learning (TL). Two images are selected from the underwater colour image dataset and those images are allowed to Discrete Wavelet Transform (DWT), Tetrolet transform and Saliency maps. Here, the outputs gained from images by the Tetrolet transform are fused and allowed for inverse Tetrolet transform. Moreover, the DWT process done with two images is fused and the output gained is allowed for inverse DWT. Similarly, the same fusion process is carried out with image outputs from Saliency maps. Finally, three image outputs that are considered as input to TL with newly devised optimization. Here, Convolutional Neural Network (CNN) is used with hyperparameters from trained models, such as SqueezeNet and AlexNet, where weights are updated using Adam Based Bald Eagle Algorithm (ABBEA). This ABBEA is obtained by combining the Bald Eagle Search (BES) algorithm and Adam Algorithm. Further, the ABBEA has Peak Signal-to-Noise Ratio (PSNR) with maximal of 38.95, Mean Squared Error (MSE) with lesser value of 20.14, Structural Similarity Index Measure (SSIM) with maximal value of 0.92, Mutual Information (MI) with maximal value of 0.86, Signal-to-Noise Ratio (SNR) with lesser value of 0.38. [ABSTRACT FROM AUTHOR]

Optical coherence tomography (OCT) is a widely used imaging technique in ophthalmology for diagnosis and treatment. Recent advances in deep neural networks (DNNs) and vision transformers (ViTs) have paved the way for automated eye/retinal disease classifications and segmentations using OCT or spectral domain OCT (SD-OCT) images. Diabetic macular edema (DME), choroidal neovascularization (CNV), and Drusen are particularly challenging to accurately classify using OCT images because of their subtle differences and intricate features. Currently, the algorithms reported in the literature using DNNs or ViTs are computationally complex, consider fewer diseases, and are less accurate. This study proposes a hybrid SqueezeNet-vision transformer (SViT) model that combines the strengths of SqueezeNet and vision transformer (ViT), capturing local and global features of OCT images to achieve more accurate classification with less computational complexity. The proposed model uses the OCT2017 dataset for training, testing, and validation, and it performs both binary classification (normal vs disorders) as well as multiclass classification (DME, CNV, Drusen, and normal). As compared to state-of-the-art CNN-based and standalone Transformer models, the proposed SViT model achieves an overall classification accuracy of 99.90% for multiclass classification (CNV: 100%, DME: 99.9%, Drusen: 100%, and normal: 100%). With a good generalization ability, the model can be used to improve patient care and clinical decision-making across a broader range of applications. [ABSTRACT FROM AUTHOR]