Layer Feedforward Research Articles

Secured Fog-Body-Torrent : A Hybrid Symmetric Cryptography with Multi-layer Feed Forward Networks Tuned Chaotic Maps for Physiological Data Transmission in Fog-BAN Environment

Recently, the Wireless Body Area Networks (WBAN) have become a promising and practical option in the tele-care medicine information system that aids for the better clinical monitoring and diagnosis. The trend of using Internet of Things (IoT) has propelled the WBAN technology to new dimension in terms of its network characteristics and efficient data transmission. However, these networks demand the strong authentication protocol to enhance the confidentiality, integrity, recoverability and dependability against the emerging cyber-physical attacks owing to the exposure of the IoT ecosystem and the confidentiality of biometric data. Hence this study proposes the Fog based WBAN infrastructure which incorporates the hybrid symmetric cryptography schemes with the chaotic maps and feed forward networks to achieve the physiological data info security without consuming the characteristics of power hungry WBAN devices. In the proposed model, scroll chaotic maps are iterated to produce the high dynamic keys streams for the real time applications and feed-forward layers are leveraged to align the complex input-output associations of cipher data for subsequent mathematical tasks. The feed forward layers are constructed which relies on the principle of Adaptive Extreme Learning Machines (AELM) thereby increasing randomness in the cipher keys thereby increasing its defensive nature against the different cyber-physical attacks and ensuring the high secured encrypted-decrypted data communication between the users and fog nodes. The real time analysis is conducted during live scenarios. BAN-IoT test beds interfaced with the heterogeneous healthcare sensors and various security metrics are analysed and compared with the various residing cryptographic algorithms. Results demonstrates that the recommended methodology has exhibited the high randomness characteristics and low computational overhead compared with the other traditional BAN oriented cryptography protocol schemes

Semantic-aware frame-event fusion based pattern recognition via large vision–language models

Pattern recognition through the fusion of RGB frames and Event streams has emerged as a novel research area in recent years. Current methods typically employ backbone networks to individually extract the features of RGB frames and event streams, and subsequently fuse these features for pattern recognition. However, we posit that these methods may suffer from two key issues: (1). They attempt to directly learn a mapping from the input vision modality to the semantic labels. This approach often leads to sub-optimal results due to the disparity between the input and semantic labels; (2). They utilize small-scale backbone networks for the extraction of RGB and Event input features, thus these models fail to harness the recent performance advancements of large-scale visual-language models. In this study, we introduce a novel pattern recognition framework that consolidates the semantic labels, RGB frames, and event streams, leveraging pre-trained large-scale vision–language models. Specifically, given the input RGB frames, event streams, and all the predefined semantic labels, we employ a pre-trained large-scale vision model (CLIP vision encoder) to extract the RGB and event features. To handle the semantic labels, we initially convert them into language descriptions through prompt engineering and polish using ChatGPT, and then obtain the semantic features using the pre-trained large-scale language model (CLIP text encoder). Subsequently, we integrate the RGB/Event features and semantic features using multimodal Transformer networks. The resulting frame and event tokens are further amplified using self-attention layers. Concurrently, we propose to enhance the interactions between text tokens and RGB/Event tokens via cross-attention. Finally, we consolidate all three modalities using self-attention and feed-forward layers for recognition. Comprehensive experiments on the HARDVS and PokerEvent datasets fully substantiate the efficacy of our proposed SAFE model. The source code has been released at https://github.com/Event-AHU/SAFE_LargeVLM.

Sky images based photovoltaic power forecasting: A novel approach with optimized VMD and Vision Mamba

As the global demand for sustainable energy sources continues to grow, accurate prediction of photovoltaic power generation is crucial for optimizing the utilization of solar resources and enhancing the efficiency of photovoltaic systems. To improve the accuracy of photovoltaic power forecasting, this paper proposes a novel hybrid predictive model that integrates Optimized Variational Mode Decomposition (VMD), Vision Mamba (Vim) for extracting features from sky images, and advanced mechanisms like Patch Embedding and Variate-wise Cross-Attention. Initially, the proposed model employs SAO-optimized VMD to decompose the photovoltaic power series into high, medium, and low-frequency components. Subsequently, these components are patched to serve as input for the subsequent layers. In the third step, exogenous variables, including meteorological and image data, are introduced and processed through Variate Embedding combined with cross-attention mechanisms to capture the intricate interactions between these variables. Finally, by integrating the outputs from all processing steps through normalization and feed-forward layers, the final predictive results are produced. Experimental evaluations across different seasons demonstrate significant enhancements in forecasting accuracy, with the model achieving Root Mean Square Error (RMSE) values of 0.3587 in spring, 0.4376 in summer, 0.3544 in autumn, and 0.3493 in winter. Similarly, Mean Absolute Error (MAE) and Mean Squared Error (MSE) across these seasons underscore the model's effectiveness. This model offers new technical means for photovoltaic power forecasting and provides valuable decision support for the optimization and management of photovoltaic power systems.

DMPR: A novel wind speed forecasting model based on optimized decomposition, multi-objective feature selection, and patch-based RNN

As the global demand for sustainable energy continues to grow, accurate wind speed prediction becomes crucial for optimizing the allocation of wind energy resources and enhancing the efficiency of wind power generation. This paper presents a composite prediction model that integrates Optimized Decomposition, Multi-Objective Feature Selection (MOFS), Patch-Based Recurrent Neural Networks (RNNs), and Cross-Attention Mechanisms. Initially, the wind speed time series is decomposed using RIME-optimized Variational Mode Decomposition (VMD), and sample entropy of each component is calculated to reconstruct high, medium, and low-frequency components. Subsequently, these components are patched to serve as inputs to the RNN layers. The third step involves the incorporation of exogenous variables, where MOFS is employed for feature selection, followed by a decomposition and reconstruction through RIME-optimized VMD. This process applies Variate-wise Cross-Attention Mechanism on the high, medium, and low-frequency components of both endogenous and exogenous variables. Finally, the outputs from all processing steps are integrated through a normalization layer and a feed-forward layer to produce the final prediction results. Experimental results demonstrate that our proposed model, through meticulous preprocessing and advanced architectural design, significantly improves the accuracy of wind speed predictions.

Assessment of blood flow parameters in a hybrid-digital model of the cardiovascular system applying recurrent neural networks

End-systolic elastance of the left ventricle along with the waveforms of pressure and volumetric blood flow in particular sectors of the circulatory system are of importance in diagnosing various problems like dilated cardiomyopathy, left-ventricular hypertrophy, pulmonary hypertension, or ischemic heart disease. The objective of the paper is to broaden the spectrum of available methods to estimate those parameters since currently accessible techniques are often costly or troublesome. Six models have been developed − three of them estimate end-systolic elastance, two perform regression of volumetric blood flow, and one predicts blood pressure. Training datasets have been collected applying the unique hybrid-digital model. The input of the designed models consists of two or three different waveforms representing pressure and volumetric blood flow in particular areas, including heart ventricles, atria, and pulmonary vessels, in addition to the heart rate value. The basis of each model comprises bidirectional Long Short-Term Memory layers along with the dropout and feed-forward layers. Models that estimate end-systolic elastance achieved various accuracy. One of them performed exceptionally well since the absolute error did not exceed 0.169 mmHgcm3 which is a negligibly small value. The root-mean-square error (RMSE) of the model predicting pressure waveform reached 0.165 mmHg in the worst case. Regression of the volumetric blood flow resulted in 6.062 cm3s worst-case RMSE for the model focusing on the pulmonary valve and 15.979 cm3s for pulmonary veins model. Computed results, especially those of the models estimating end-systolic elastance, indicate that it is possible to utilize neural networks to estimate those parameters with sufficient accuracy.

Rethinking encoder-decoder architecture using vision transformer for colorectal polyp and surgical instruments segmentation

Accurate polyp segmentation from colonoscopy images is important for the immediate diagnosis and effective treatment of colon cancer. While significant progress has been made in the polyps segmentation task, there are various challenges that need to be addressed. Polyps can vary greatly in size and shape, and often has no clear boundary between surrounding tissues. Furthermore, surgical instrument segmentation can aid surgeons with the precise positioning and orientation of the instruments, helping them to plan the next steps in the robot-assisted surgery. The proposed colorectal polyp segmentation Transformer (CPS-Former) uses innovative attention blocks in a network that encodes-decodes features like classic Semantic Segmentation Network (SegNet). However, it has special self-attention modules with small convolutional kernels that efficiently extract information from different feature-channels. Moreover, it is equipped with an effective positional embedding to capture information from a large area of context for long distance interactions. Additionally, a fusion block is embedded for scaling-attention that combines the outputs from the encoder-decoder blocks to enhance the semantic features and reduce the non-semantic ones. Transformer encoder blocks also modified by adding a local feedforward layer and skips connections, and adjust the channel sizes to reduce the model trainable parameters. We evaluate our colorectal polyp segmentation network (CPS-Former) on four colorectal polyp public datasets and one surgical instrument segmentation dataset, which show its superiority over other state-of-the-art polyp segmentation models. Our implementation source code and network weights are available at GitHub: https://github.com/ahmedeqbal/CPS-Former.

Stereoscopic video deblurring transformer

Stereoscopic cameras, such as those in mobile phones and various recent intelligent systems, are becoming increasingly common. Multiple variables can impact the stereo video quality, e.g., blur distortion due to camera/object movement. Monocular image/video deblurring is a mature research field, while there is limited research on stereoscopic content deblurring. This paper introduces a new Transformer-based stereo video deblurring framework with two crucial new parts: a self-attention layer and a feed-forward layer that realizes and aligns the correlation among various video frames. The traditional fully connected (FC) self-attention layer fails to utilize data locality effectively, as it depends on linear layers for calculating attention maps The Vision Transformer, on the other hand, also has this limitation, as it takes image patches as inputs to model global spatial information. 3D convolutional neural networks (3D CNNs) process successive frames to correct motion blur in the stereo video. Besides, our method uses other stereo-viewpoint information to assist deblurring. The parallax attention module (PAM) is significantly improved to combine the stereo and cross-view information for more deblurring. An extensive ablation study validates that our method efficiently deblurs the stereo videos based on the experiments on two publicly available stereo video datasets. Experimental results of our approach demonstrate state-of-the-art performance compared to the image and video deblurring techniques by a large margin.

Precision forecasting of grinding wheel Wear: A TransBiGRU model for advanced industrial predictive maintenance

In intelligent manufacturing, accurate prediction of grinding wheel wear is essential to reduce maintenance costs and improve production efficiency. To achieve accurate forecasts, this paper introduces a grinding wheel wear prediction model, TransBiGRU, incorporating a Transformer encoder, Bidirectional Gated Recurrent Unit (BiGRU), positional encoding layer, and position-wise feedforward layer. The model extracts features by analyzing current signal characteristics in basic statistical, time, and frequency domains during the machining process. Training is conducted through K-fold cross-validation to ensure model stability. The experimental results indicate that the model achieved good performance with Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Determination (R2) evaluation metrics, obtaining values of 2.0898, 3.262, and 0.9338, respectively. Systematically reducing modules validates the importance of each module in enhancing predictive capabilities. This research achieves the practical application of low-cost current sensors in optimizing maintenance plans and reducing downtime in predicting grinding wheel wear.

DeepChestGNN: A Comprehensive Framework for Enhanced Lung Disease Identification through Advanced Graphical Deep Features.

Lung diseases are the third-leading cause of mortality in the world. Due to compromised lung function, respiratory difficulties, and physiological complications, lung disease brought on by toxic substances, pollution, infections, or smoking results in millions of deaths every year. Chest X-ray images pose a challenge for classification due to their visual similarity, leading to confusion among radiologists. To imitate those issues, we created an automated system with a large data hub that contains 17 datasets of chest X-ray images for a total of 71,096, and we aim to classify ten different disease classes. For combining various resources, our large datasets contain noise and annotations, class imbalances, data redundancy, etc. We conducted several image pre-processing techniques to eliminate noise and artifacts from images, such as resizing, de-annotation, CLAHE, and filtering. The elastic deformation augmentation technique also generates a balanced dataset. Then, we developed DeepChestGNN, a novel medical image classification model utilizing a deep convolutional neural network (DCNN) to extract 100 significant deep features indicative of various lung diseases. This model, incorporating Batch Normalization, MaxPooling, and Dropout layers, achieved a remarkable 99.74% accuracy in extensive trials. By combining graph neural networks (GNNs) with feedforward layers, the architecture is very flexible when it comes to working with graph data for accurate lung disease classification. This study highlights the significant impact of combining advanced research with clinical application potential in diagnosing lung diseases, providing an optimal framework for precise and efficient disease identification and classification.

Stress-Nets- A Novel LSTM Ensembled Single Feed Forward Layers for Stress Classification with EEG Signals

Mental instability and emotional imbalance of the individual can be reflected in the form of stress which results in an inappropriate work ethics. There are various methods for the Stress creation. Moreover , several bio-signal sources such as Electroencephalograph(EEG), Electrocardiography(ECG) and Electromyography(EMG) are considered to be the important catalyst for developing the stress detection systems(SDS). Recently, machine and deep learning algorithms has gained more popularity in designing the SDS using the bio-signals. Further, EEG based SDS with ML and DL algorithms plays an important role for its high classification accuracy and performance. However, these EEG based SDS systems needs the better lime light of improvisation in terms of performance and computational overhead to deal with multiple datasets. In this context, this manuscript proposes new Long Short term Memory recurrent units(LSTM) for extracting temporal features while enhanced extreme learning machines are employed for better classification with less computational complexity. The different data sources are used to collect the EEG signals in which the collected signals are preprocessed for evaluating the proposed model. Additionally, the experiments are performed by DEAP and Kaggle datasets as well as performance parameters and compared by conventional Fused Support vector machines (F-SVM),BI-Long Short Term Memory(BILSTM), Random Forest(RF) and Deep Convolutional Neural network(DCNN). Results shows which proposed EEG based SDS has better performance than other conventional ones of high accuracy in stress detection and for diagnosing classify the stress-levels .

GRAformer: A gated residual attention transformer for multivariate time series forecasting

Recurrent Neural Networks (RNNs), particularly when equipped with output windows – a standard practice in contemporary time series forecasting – have shown proficiency in handling short-term dependencies. Nonetheless, RNNs can encounter challenges in maintaining hidden states over extended forecasting periods, particularly in longer-term predictions where increased hidden state sizes and extended look-back windows can lead to gradient instability. In contrast, Transformer-based models, with their distinctive architecture designed to encode complex contextual relationships and enable computations to be done in parallel, are emerging as a popular alternative in this field. However, current research has mainly focused on modifying attention mechanisms, overlooking opportunities to improve the feedforward layer, which could lead to efficiency limitations. Moreover, prevailing methods often assume absolute independence between channels, disregarding distinct features among variables and failing to fully leverage channel-specific information. To address these gaps, we propose an efficient transformer design for multivariate time series prediction. Our approach integrates two key components: (i) a gated residual attention unit that enhances predictive accuracy and computational efficiency, and (ii) a channel embedding technique that differentiates between series and boosts performance. Theoretically, we prove that our model has recurrent dynamics introduced by the RNN layer. Through extensive experiments on real-world data, we demonstrate that our proposed method achieves competitive predictive accuracy compared to prior approaches while exhibiting accelerated processing relative to state-of-the-art transformers. Our code, data, and trained models are available at https://github.com/MythosAd/GRAformer.

A deep learning approach for balance optimisation of patch panel assembly line

This paper introduces an integrated methodology to enhance the efficiency of patch panel assembly lines by harnessing advanced artificial intelligence (AI) techniques. The primary objective is to minimise the total completion time of patch panels, a pivotal metric for manufacturing efficiency. The study employs the Assembly Line Production Optimisation (ALPO) model to analyze and enhance the production capacity of the bottleneck station. Existing challenges related to patch panel assembly line task distribution, process content, and tooling layout are scrutinised against production process standards for operational measurements. To achieve improvement, Deep Neural Network (DNN) and Recurrent Neural Network (RNN) models are employed in hidden layers, supplemented by deep feed-forward (DFF) layers for output, in adherence to engineering design principles. The bottleneck station's load rate is successfully reduced to below 80%, contributing to an overall balance rate increase from 73% to 88%. These outcomes signify a notable enhancement in production efficiency and employee satisfaction. The research utilises a dataset derived from the Microsoft Azure AI-based PdM dataset. The paper illustrates how AI techniques offer a comprehensive solution to address the inherent complexities in patch panel assembly line optimisation.

Stage-Wise Magnitude-Based Pruning for Recurrent Neural Networks.

A recurrent neural network (RNN) has shown powerful performance in tackling various natural language processing (NLP) tasks, resulting in numerous powerful models containing both RNN neurons and feedforward neurons. On the other hand, the deep structure of RNN has heavily restricted its implementation on mobile devices, where quite a few applications involve NLP tasks. Magnitude-based pruning (MP) is a promising way to address such a challenge. However, the existing MP methods are mostly designed for feedforward neural networks that do not involve a recurrent structure, and, thus, have performed less satisfactorily on pruning models containing RNN layers. In this article, a novel stage-wise MP method is proposed by explicitly taking the featured recurrent structure of RNN into account, which can effectively prune feedforward layers and RNN layers, simultaneously. The connections of neural networks are first grouped into three types according to how they are intersected with recurrent neurons. Then, an optimization-based pruning method is applied to compress each group of connections, respectively. Empirical studies show that the proposed method performs significantly better than the commonly used RNN pruning methods; i.e., up to 96.84% connections are pruned with little or even no degradation of precision indicators on the testing datasets.

Digitized Perception of Piano Accompaniment in Songs with the Involvement of Deep Learning Techniques

Abstract In this paper, we use the Hanning window function to add windows to the piano accompaniment human brain perception signal, and the noise and redundant features in the piano accompaniment human brain perception signal are downgraded by principal component analysis and Relief algorithm for dimensionality reduction processing and feature selection. Based on the basic structure of RCNN, the ReConv-EEG Net EEG identity recognition method is established, and the feature vectors in the recurrent network are classified and computed through the feed-forward layer, the normalization layer and the output layer to obtain the feature values of the EEG signals. Under different speed accompaniments, the subjects perceived the high-speed piano accompaniment with the lowest α -wave power spectral value of 47.85 on average, while the original speed accompaniment had the most significant α -wave power spectral value of 49.12 on average.Meanwhile, there were differences in the perception of piano accompaniment among subjects, and the mean EEG δ -wave synchronization indexes of the non-musicians and the musicians differed by an average of 0.2297 when they were listening to the same accompaniment.The method in this paper contributes to developing a new method for deep learning-assisted music perception research in the field. Also, it lays the data foundation for applying the technique to the field.

TTST: A Top-k Token Selective Transformer for Remote Sensing Image Super-Resolution.

Transformer-based method has demonstrated promising performance in image super-resolution tasks, due to its long-range and global aggregation capability. However, the existing Transformer brings two critical challenges for applying it in large-area earth observation scenes: (1) redundant token representation due to most irrelevant tokens; (2) single-scale representation which ignores scale correlation modeling of similar ground observation targets. To this end, this paper proposes to adaptively eliminate the interference of irreverent tokens for a more compact self-attention calculation. Specifically, we devise a Residual Token Selective Group (RTSG) to grasp the most crucial token by dynamically selecting the top- k keys in terms of score ranking for each query. For better feature aggregation, a Multi-scale Feed-forward Layer (MFL) is developed to generate an enriched representation of multi-scale feature mixtures during feed-forward process. Moreover, we also proposed a Global Context Attention (GCA) to fully explore the most informative components, thus introducing more inductive bias to the RTSG for an accurate reconstruction. In particular, multiple cascaded RTSGs form our final Top- k Token Selective Transformer (TTST) to achieve progressive representation. Extensive experiments on simulated and real-world remote sensing datasets demonstrate our TTST could perform favorably against state-of-the-art CNN-based and Transformer-based methods, both qualitatively and quantitatively. In brief, TTST outperforms the state-of-the-art approach (HAT-L) in terms of PSNR by 0.14 dB on average, but only accounts for 47.26% and 46.97% of its computational cost and parameters. The code and pre-trained TTST will be available at https://github.com/XY-boy/TTST for validation.

Transformer-based cascade networks with spatial and channel reconstruction convolution for deepfake detection.

The threat posed by forged video technology has gradually grown to include individuals, society, and the nation. The technology behind fake videos is getting more advanced and modern. Fake videos are appearing everywhere on the internet. Consequently, addressing the challenge posed by frequent updates in various deepfake detection models is imperative. The substantial volume of data essential for their training adds to this urgency. For the deepfake detection problem, we suggest a cascade network based on spatial and channel reconstruction convolution (SCConv) and vision transformer. Our network model's front portion, which uses SCConv and regular convolution to detect fake videos in conjunction with vision transformer, comprises these two types of convolution. We enhance the feed-forward layer of the vision transformer, which can increase detection accuracy while lowering the model's computing burden. We processed the dataset by splitting frames and extracting faces to obtain many images of real and fake faces. Examinations conducted on the DFDC, FaceForensics++, and Celeb-DF datasets resulted in accuracies of 87.92, 99.23 and 99.98%, respectively. Finally, the video was tested for authenticity and good results were obtained, including excellent visualization results. Numerous studies also confirm the efficacy of the model presented in this study.

GRU- and Transformer-Based Periodicity Fusion Network for Traffic Forecasting

Accurate traffic prediction is vital for traffic management, control, and urbanization construction. Extensive research efforts have diligently focused on capturing the intricate spatio-temporal relationships that are inherent in traffic data. However, a limited number of studies have fully exploited the potential of periodicity, a distinctive and valuable characteristic of transportation systems. In this paper, we propose a novel GRU- and Transformer-Based Periodicity Fusion Network (GTPFN) to distinguish the effects of different types of periodic data and integrate them seamlessly and effectively. Initially, the proposed model captures dynamic spatio-temporal correlations and obtains the candidate prediction result by employing a GRU encoder–decoder with spatial attention, focusing on the hourly data. Subsequently, we design the Pattern Induction Block based on GRU layers to extract regular traffic patterns from daily and weekly data. Finally, the Pattern Fusion Transformer integrates these patterns, followed by a Feedforward layer, to yield the final prediction output. Experiments on the Caltrans Performance Measurement System (PEMS) datasets illustrate that the proposed model outperforms state-of-art baseline models on most predicted horizons.

Learning heuristics for arc routing problems

Solid waste collection and street sweeping can each be formulated as a capacitated arc routing problem (CARP) in which the streets of a city require service by a fleet of vehicles at minimum total cost. Heuristic and meta-heuristic methods have been used to solve CARP because the problem is known to be NP-Hard. In this paper we introduce Splice, a novel machine learning framework that learns heuristics to solve routing problems, such as CARP, defined on the edges (rather than vertices) of a non-euclidean graph representation of the problem. Splice uses a message passing graph neural network to generate a graph embedding, and deep Q-learning feed-forward layers to learn heuristics to construct a giant tour that results in minimum total cost after applying the splitting procedure, iterating until all tasks are serviced. Compared to the two AI-driven approaches to CARP, Splice has a simpler network architecture and does not require a supervised learning method for training. The total costs of the solutions found by Splice are from 11-30% lower than those found by a path scanning heuristic and a memetic meta-heuristic algorithm. When trained on small instances, the heuristics learned by Splice generalize well, i.e., can be used to solve larger instances without retraining. Due to the route-first split-second form of heuristic learned, the Splice framework may be applied to solve other variants of routing problems by adapting the splitting procedure.

Dual-modality learning and transformer-based approach for high-quality vector font generation

Vector fonts, serving as the fundamental format of fonts, play a significant role in modern media. Described by a set of mathematical equations, vector fonts enable style modifications by adjusting drawing parameters, making them favored by font artists and designers. Due to the non-structural nature of vector font data, the task of vector font generation resembles sequence generation. Existing methods, limited in handling long sequences, are only capable of synthesizing simple character vector fonts. In this paper, we propose a dual-modal learning strategy to convert raster glyph images to vector glyph images in an end-to-end manner. Specifically, by employing vector quantization, we comprehensively utilize the dual-modal information of vector fonts. We quantize image and sequence modal features using a shared codebook, mapping them to the same discrete space and aligning them. Through aligned features, we reconstruct raster glyph images and vector glyph images. During the transformation process of vector glyph data, we redesign the Transformer module for vector glyph data, leveraging multiple stacked sliding window attention mechanisms to model local and global information. By integrating reversible residuals with attention and feedforward layers within the Transformer module, we enhance the model’s capability and stability in handling long sequence data without sacrificing accuracy. Finally, we perform cross-modal model distillation to obtain the model’s backbone network. We further refine the backbone network using a differentiable rasterizer to minimize error accumulation during sequence generation. Qualitative and quantitative results demonstrate that our method achieves high-quality synthesis results in complex Chinese character glyphs. The synthesized vector fonts can be easily converted into TrueType fonts for practical use, holding valuable applications for anyone interested in personalized vector font styles.

FLY-CAPS- A Hybrid Firefly Feature Optimized Capsule Networks for Plant Disease Classification in Resource Constriant Internet of Things (IoT)

Recent advancements in artificial intelligence, automation, and the Internet of Things (IoT) enable farmers to better monitor and diagnose all agricultural procedures with super-intellectual accuracy. These technologies also contribute to boosting the productivity of agriculture, which increases the country’s economy. Though these technologies help farmers increase productivity, the detection of plant diseases still needs heightened scrutiny for prevention and cultivation. Plant disease categorization has expanded with the introduction of deep learning algorithms, but it still needs more innovation in terms of accuracy and computing burden. Thus, a novel deep learning model based on capsule networks with firefly optimization and potent multi-layered feedforward prediction networks is proposed in this research. The handcrafted features in this proposed system are optimized before being extracted using a capsule network, which reduces the complexity overhead and is suitable for IoT devices with limited resources. Finally fed to the feed forward layers for better classification. The extensive experimentation has been tested with the Plant Village databases, which contain more than 50,000 images of healthy and infected plants. Performance criteria including recall, specificity, recall, accuracy, and f1-score are used to assess the proposed algorithm's performance. Additionally, its efficiency and computational cost are contrasted with those of other recent models. The suggested model has greater performance (95%) with reduced computing overhead, according to experimental data, which is advantageous for the new prediction approach and the welfare of the farmer.