Pre-trained Convolutional Neural Network Model Research Articles

Innovative Deep Learning Architecture for the Classification of Lung and Colon Cancer From Histopathology Images

The increasing prevalence of colon and lung cancer presents a considerable challenge to healthcare systems worldwide, emphasizing the critical necessity for early and accurate diagnosis to enhance patient outcomes. The precision of diagnosis heavily relies on the expertise of histopathologists, constituting a demanding task. The health and well‐being of patients are jeopardized in the absence of adequately trained histopathologists, potentially leading to misdiagnoses, unnecessary treatments, and tests, resulting in the inefficient utilization of healthcare resources. However, with substantial technological advancements, deep learning (DL) has emerged as a potent tool in clinical settings, particularly in the realm of medical imaging. This study leveraged the LC25000 dataset, encompassing 25,000 images of lung and colon tissue, introducing an innovative approach by employing a self‐organized operational neural network (Self‐ONN) to accurately detect lung and colon cancer in histopathology images. Subsequently, our novel model underwent comparison with five pretrained convolutional neural network (CNN) models: MobileNetV2‐SelfMLP, Resnet18‐SelfMLP, DenseNet201‐SelfMLP, InceptionV3‐SelfMLP, and MobileViTv2_200‐SelfMLP, where each multilayer perceptron (MLP) was replaced with Self‐MLP. The models’ performance was meticulously assessed using key metrics such as precision, recall, F1 score, accuracy, and area under the receiver operating characteristic (ROC) curve. The proposed model demonstrated exceptional overall accuracy, precision, sensitivity, F1 score, and specificity, achieving 99.74%, 99.74%, 99.74%, 99.74%, and 99.94%, respectively. This underscores the potential of artificial intelligence (AI) to significantly enhance diagnostic precision within clinical settings, portraying a promising avenue for improving patient care and outcomes. The synopsis of the literature provides a thorough examination of several DL and digital image processing methods used in the identification of cancer, with a primary emphasis on lung and colon cancer. The experiments use the LC25000 dataset, which consists of 25,000 photos, for the purposes of training and testing. Various techniques, such as CNNs, transfer learning, ensemble models, and lightweight DL architectures, have been used to accomplish accurate categorization of cancer tissue. Various investigations regularly show exceptional performance, with accuracy rates ranging from 96.19% to 99.97%. DL models such as EfficientNetV2, DHS‐CapsNet, and CNN‐based architectures such as VGG16 and GoogleNet variations have shown remarkable performance in obtaining high levels of accuracy. In addition, methods such as SSL and lightweight DL models provide encouraging outcomes in effectively managing large datasets. In general, the research emphasizes the efficacy of DL methods in successfully diagnosing cancer from histopathological pictures. It therefore indicates that DL has the potential to greatly improve medical diagnostic techniques.

Automatic classification of 10 blood cell subtypes using transfer learning via pre-trained convolutional neural networks

Human blood is primarily composed of plasma, red blood cells, white blood cells, and platelets. It plays a vital role in transporting oxygen and nutrients to all organs, and stores essential health-related data about the human body. Blood cells are utilized to defend the body against infections and disease. Hence, analysis of blood permits physicians to assess an individual's physiological condition. Blood cells are sub-classified into eight groups: Neutrophils, eosinophils, basophils, lymphocytes, monocytes, immature granulocytes (promyelocytes, myelocytes, and metamyelocytes), erythroblasts, and platelets (thrombocytes) on the basis of their nucleus, shape and cytoplasm. Traditionally, pathologists and hematologists have identified and examined these via microscopy prior to manual classification, with this manual approach being slow and prone to human error. Therefore, it is essential to automate this process. In the current study, transfer learning with a series of pre-trained Convolutional Neural Network (CNN) models—VGG16, VGG19, ResNet-50, ResNet-101, ResNet-152, InceptionV3, MobileNetV2 and DenseNet-201 was applied to the normal peripheral blood cells dataset (PBC). The overall accuracy achieved with individual CNNs ranged from 91.4 % to 94.7 %. Based on these pre-trained architectures, a CNN-based architecture has been developed to automatically classify all ten blood cell types. The proposed transfer learning CNN model was tested on blood cell images from the PBC, Kaggle and LISC datasets. Achieved accuracy was 99.91 %, 99.68 % and 98.79 %, respectively, across these three datasets. The presented CNN architecture outperforms all previous results reported in the scientific/medical literature with a high capacity for framework generalization in future applications of blood cell classification.

Convolutional Neural Network-Enhanced Video Artistry: Leveraging VGG16 for Dynamic Style Transfer

This paper introduces a novel application of neural style transfer to video sequences, leveraging pre-trained Convolutional Neural Network (CNN) models. By combining content and style losses in an optimization process, the approach transforms video frames into visually captivating compositions, opening new avenues for artistic expression in visual storytelling and filmmaking. The procedure begins with meticulous preprocessing of video frames to align them with the specifications of the powerful VGG16 model. A content model is then constructed to extract activations from the 14th layer of the VGG network, forming the groundwork for nuanced content depiction. Style information is computed by aggregating data from a combination of convolutional layers, resulting in a multifaceted, artistically complex tapestry. The paper introduces a well-balanced loss function that combines style and content losses, serving as the foundation for the ensuing optimization procedure. This careful balancing between style adherence and content authenticity yields visually stunning synthesis, enhancing video sequence quality and artistic resonance. The paper also delves into a detailed analysis of hyper parameters, particularly style weights, and introduces a novel pixel value scaling method for enhanced visual coherence. Experimental findings demonstrate the efficiency and adaptability of the proposed approach, showcasing the seamless integration of content and style for an immersive visual experience. This accomplishment marks a significant stride toward unlocking the creative potential of video content, ushering in a period of enhanced visual storytelling.

Evaluation of Deep Learning Models in Search by Example using Capsule Endoscopy Images

Wireless capsule endoscopy (WCE) has revolutionized the field of gastrointestinal examinations, being MedtronicTM WCE one of the most used in clinics. In those WCE videos, medical experts use RAPID READERTM tool to annotate findings in videos. However, the frame annotations are not available in an open format and, when exported, they have different resolutions and some annotated artefacts that make difficult their localization in the original videos. This difficult the use of WCE medical experts’ annotations in the research of new computed-aid diagnostic (CAD) methods. In this paper, we propose a methodology to compare image similarities and evaluate it in a private MedtronicTM WCE SB3 video dataset to automatically identify the annotated frames in the videos. We used state-of-the-art pre-trained convolutional neural network (CNN) models, including MobileNet, InceptionResNetv2, ResNet50v2, VGG19, VGG16, ResNet101v2, ResNet152v2, and DenseNet121, as frame features extractors and compared them with the Euclidean distance. We evaluated the methodology performance on a private dataset consisting of 100 WCE videos, totalling 905 frames. The experimental results showed promising performance. The MobileNet model achieved an accuracy of 94% for identifying the first match, while the top 5, top 10, and top 20 matches were identified with accuracies of 94%, 94%, and 98%, respectively. The VGG16 and ResNet50v2 models also demonstrated strong performance, achieving accuracies ranging from 88% to 93% for various match positions. These results highlight the effectiveness of our proposed methodology in localizing target frames and even identifying similar frames very use useful for training data-driven models in CAD research. The code utilized in this experiment is available on the Github†.

Car make and model recognition using convolutional neural network: fine-tune AlexNet architecture

Artificial intelligence (AI) has significantly contributed to car make and model recognition in this current era of intelligent technology. By using AI, it is much easier to identify car models from any picture or video. This paper introduces a new model by fine-tuning the AlexNet architecture to determine the car model from images. First of all, our car image dataset has been created. Some of these images were taken by us, and others were taken from the website of the car connection. Then we cleaned all the unwanted images for better performance. Our dataset has ten classes containing 5,000 car images split into train and test data. After that, we augmented our data with random flip, rotation, and zoom to reduce overfitting. Finally, we used a pre-trained convolutional neural network (CNN) model AlexNet architecture. We fine-tuned AlexNet (FT-AlexNet) by adding three extra layers for better classification and compared it with the original AlexNet. To measure the performance of these models, accuracy, precision, recall, and F1-score were used. The results show that fine-tune AlexNet architecture outperforms the original AlexNet architecture. The results prove that recognition accuracy has increased due to our improvement approach.

Evaluation of Deep Learning Models in Search by Example using Capsule Endoscopy Images

Wireless capsule endoscopy (WCE) has revolutionized the field of gastrointestinal examinations, being MedtronicTM WCE one of the most used in clinics. In those WCE videos, medical experts use RAPID READERTM tool to annotate findings in videos. However, the frame annotations are not available in an open format and, when exported, they have different resolutions and some annotated artefacts that make difficult their localization in the original videos. This difficult the use of WCE medical experts’ annotations in the research of new computed-aid diagnostic (CAD) methods. In this paper, we propose a methodology to compare image similarities and evaluate it in a private MedtronicTM WCE SB3 video dataset to automatically identify the annotated frames in the videos. We used state-of-the-art pre-trained convolutional neural network (CNN) models, including MobileNet, InceptionResNetv2, ResNet50v2, VGG19, VGG16, ResNet101v2, ResNet152v2, and DenseNet121, as frame features extractors and compared them with the Euclidean distance. We evaluated the methodology performance on a private dataset consisting of 100 WCE videos, totalling 905 frames. The experimental results showed promising performance. The MobileNet model achieved an accuracy of 94% for identifying the first match, while the top 5, top 10, and top 20 matches were identified with accuracies of 94%, 94%, and 98%, respectively. The VGG16 and ResNet50v2 models also demonstrated strong performance, achieving accuracies ranging from 88% to 93% for various match positions. These results highlight the effectiveness of our proposed methodology in localizing target frames and even identifying similar frames very use useful for training data-driven models in CAD research. The code utilized in this experiment is available on the Github†.

A data-driven fusion of deep learning and transfer learning for orange disease classification

In agriculture, early detection of crop diseases is imperative for sustainability and maximizing yields. Rooted in Agriculture 4.0, our innovative approach combines pre-trained Convolutional Neural Networks (CNNs) models with data-driven solutions to address global challenges related to water scarcity. By integrating the combined L1/L2 regularization technique to our model layers, we enhance their flexibility, reducing the risk of the overfitting effect of the model. In the orange dataset used in our experiments, we have 1790 orange images, including a class of fresh oranges and three disease categories. Applied on this dataset for classification, our model exhibits notable performance, namely 92.17% for CNN and 97.28% for ResNet-50 model. Evaluated across metrics like accuracy, precision, recall, F1-score, confusion matrix, and cross validation, our approach surpasses traditional classifiers, significantly contributing to smart agricultural and global food resilience amidst mounting water scarcity pressures.

Locating Target Regions for Image Retrieval in an Unsupervised Manner.

Image retrieval performance can be improved by training a convolutional neural network (CNN) model with annotated data to facilitate accurate localization of target regions. However, obtaining sufficiently annotated data is expensive and impractical in real settings. It is challenging to achieve accurate localization of target regions in an unsupervised manner. To address this problem, we propose a new unsupervised image retrieval method named unsupervised target region localization (UTRL) descriptors. It can precisely locate target regions without supervisory information or learning. Our method contains three highlights: 1) we propose a novel zero-label transfer learning method to address the problem of co-localization in target regions. This enhances the potential localization ability of pretrained CNN models through a zero-label data-driven approach; 2) we propose a multiscale attention accumulation method to accurately extract distinguishable target features. It distinguishes the importance of features by using local Gaussian weights; and 3) we propose a simple yet effective method to reduce vector dimensionality, named twice-PCA-whitening (TPW), which reduces the performance degradation caused by feature compression. Notably, TPW is a robust and general method that can be widely applied to image retrieval tasks to improve retrieval performance. This work also facilitates the development of image retrieval based on short vector features. Extensive experiments on six popular benchmark datasets demonstrate that our method achieves about 7% greater mean average precision (mAP) compared to existing state-of-the-art unsupervised methods.

Pre-trained CNNs: Evaluating Emergency Vehicle Image Classification

In this paper, we aim to provide a comprehensive analysis of image classification, specifically in the context of emergency vehicle classification. We have conducted an in-depth investigation, exploring the effectiveness of six pre-trained Convolutional Neural Network (CNN) models. These models, namely VGG19, VGG16, MobileNetV3Large, MobileNetV3Small, MobileNetV2, and MobileNetV1, have been thoroughly examined and evaluated within the domain of emergency vehicle classification. The research methodology utilized in this study is carefully designed with a systematic approach. It includes the thorough preparation of datasets, deliberate modifications to the model architecture, careful selection of layer operations, and fine-tuning of the model compilation. To gain a comprehensive understanding of the performance, we conducted a detailed series of experiments. We analyzed nuanced performance metrics such as accuracy, loss, and training time, considering important factors in the evaluation process. The results obtained from this study provide a comprehensive understanding of the advantages and disadvantages of each model. Moreover, they emphasize the crucial significance of carefully choosing a suitable pre-trained Convolutional Neural Network (CNN) model for image classification tasks. Essentially, this article provides a comprehensive overview of image classification, highlighting the crucial significance of pre-trained CNN models in achieving precise outcomes, especially in the demanding field of emergency vehicle classification.

Deep Learning-Based Fire Detection for Enhanced Safety Systems

Fire detection systems are a critical aspect of modern safety and security systems, playing a pivotal role in safeguarding lives and property against the destructive force of fires. Rapid and accurate identification of fire incidents is essential for timely response and mitigation efforts. Traditional fire detection methods have made substantial advancements, but with the advent of computer vision technologies, the field has witnessed a transformative shift. This paper presents a method for fire detection using deep convolutional neural network (CNN) models. This approach used transfer learning by employing two pre-trained CNN models from the ImageNet dataset: VGG (Visual Geometry Group) and InceptionV3 to extract valuable features from input images. Then, these extracted features serve as input for a machine learning (ML) classifier, namely the Softmax classifier. The Softmax activation function computes the probability distribution to assign accurate class probabilities for discriminating between two types of images: fire and non-fire. Experimental results showed that the proposed method successfully detected fire areas and achieved seamless classification performance compared to other current fire detection methods.

An Evaluation of Pre-Trained Convolutional Neural Network Models for the Detection of COVID-19 and Pneumonia from Chest X-Ray Imagery

An Evaluation of Pre-Trained Convolutional Neural Network Models for the Detection of COVID-19 and Pneumonia from Chest X-Ray Imagery

Image classification of lotus in Nong Han Chaloem Phrakiat Lotus Park using convolutional neural networks

The Nong Han Chaloem Phrakiat Lotus Park is a tourist attraction and a source of learning regarding lotus plants. However, as a training area, it lacks appeal and learning motivation due to its conventional presentation of information regarding lotus plants. The current study introduced the concept of smart learning in this setting to increase interest and motivation for learning. Convolutional neural networks (CNNs) were used for the classification of lotus plant species, for use in the development of a mobile application to display details about each species. The scope of the study was to classify 11 species of lotus plants using the proposed CNN model based on different techniques (augmentation, dropout, and L2) and hyper parameters (dropout and epoch number). The expected outcome was to obtain a high-performance CNN model with reduced total parameters compared to using three different pre-trained CNN models (Inception V3, VGG16, and VGG19) as benchmarks. The performance of the model was presented in terms of accuracy, F1-score, precision, and recall values. The results showed that the CNN model with the augmentation, dropout, and L2 techniques at a dropout value of 0.4 and an epoch number of 30 provided the highest testing accuracy of 0.9954. The best proposed model was more accurate than the pre-trained CNN models, especially compared to Inception V3. In addition, the number of total parameters was reduced by approximately 1.80–2.19 times. These findings demonstrated that the proposed model with a small number of total parameters had a satisfactory degree of classification accuracy.

A generalization ability-enhanced image recognition based multiaxial fatigue life prediction method for complex loading conditions

A generalization ability-enhanced image recognition based multiaxial low-cycle fatigue life prediction is proposed for complex conditions, including multiaxial variable amplitude fatigue, fatigue-creep interaction, and anisotropic materials. Notably, a novel modified dynamic hysteresis image plotting method is introduced for multiaxial variable amplitude loading. Building on a pre-trained convolution neural network (CNN) model, fine-tuning is used across three cases. It makes the error decrease to 25%. Besides, fine-tuning improves training efficiency and generalization. Based on fine-tuning results, each layer of the model is attributed with physics significance. Furthermore, regression activation map (RAM) is used for interpretation, illustrating that the model focuses on the inner and edges of hysteresis images.

Two-Stage Input-Space Image Augmentation and Interpretable Technique for Accurate and Explainable Skin Cancer Diagnosis

This research paper presents a deep-learning approach to early detection of skin cancer using image augmentation techniques. We introduce a two-stage image augmentation process utilizing geometric augmentation and a generative adversarial network (GAN) to differentiate skin cancer categories. The public HAM10000 dataset was used to test how well the proposed model worked. Various pre-trained convolutional neural network (CNN) models, including Xception, Inceptionv3, Resnet152v2, EfficientnetB7, InceptionresnetV2, and VGG19, were employed. Our approach demonstrates an accuracy of 96.90%, precision of 97.07%, recall of 96.87%, and F1-score of 96.97%, surpassing the performance of other state-of-the-art methods. The paper also discusses the use of Shapley Additive Explanations (SHAP), an interpretable technique for skin cancer diagnosis, which can help clinicians understand the reasoning behind the diagnosis and improve trust in the system. Overall, the proposed method presents a promising approach to automated skin cancer detection that could improve patient outcomes and reduce healthcare costs.

AutomatedFace Mask Detection Using Pretrained CNN

In recent times, the use of face masks has emerged as a critical subject. Automated facial mask detection can curb the transmission of the COVID-19 virus and SARS-VIRUS within communal areas by identifying individuals who are not utilizing masks. In this work, a pretrained Convolutional Neural Network (CNN) with model ResNet-50, which is initially trained on the Image Net competition data, is utilized. This model is augmented with a 300-linear layer network and fine-tuned on a dataset that comprises 1,000 facial images. During the evaluation of the validation dataset consisting of approximately 800 face images, the model achieved an impressive 99% accuracy. Its primary objective is to discover if an individual is wearing a facial mask using a cropped image of their face. By leveraging such advanced technologies, we can contribute significantly to public health and safety measures in the ongoing battle against COVID-19 and SARS-VIRUS.Keywords:CNNCOVID-19Face Mask Detection (FMD) SARSDOI: 10.22401/ANJS.26.4.12*corresponding author email: farah.saad@nahrainuniv.edu.iqThis work is licensed under a Creative Commons Attribution 4.0 International License1.Introduction COVID-19 (CO ronaVIrus Disease of 2019) and SARS(Severe Acute Respiratory Syndrome) are two viral respiratory illnesses caused by coronaviruses, both of which have had important global impacts on public health [1]. These contagious diseases emerged in different time frames but share similarities in their mode of transmission and clinical presentation, prompting the search for effective non-pharmaceutical strategies to mitigate their spread [2]. One essential strategy to reduce the transmission of respiratory droplets, which can carry the viruses, is the widespread adoption of face masks in public spaces. However, monitoring and ensuring compliance with mask-wearing practices in crowded areas can be challenging, particularly in densely populated regions. Automatic face mask detection (FMD) systems powered by advanced technologies offer a potential solution to this challenge [3]. Convolutional Neural Networks (CNNs) and deep learning techniques have demonstrated remarkable success in computer vision tasks, including image classification [4]. In this context, utilizing pretrained CNN models, such as ResNet-50, has shown promising results for identifying faces wearing or not wearing masks. In this work, we explore applying a pretrained ResNet-50 CNN model fine-tuned on a well-balanced dataset of 12,000cropped photos of faces with and without masks. The goal is to create an accurate and reliable FMD (Face Mask Detection) system that can automatically identify individuals who are not wearing masks in public spaces. Through this work, we aim to contribute to the expanding realm of research on FMD as a crucial tool to prevent the spread of respiratory illnesses. The successful implementation of such technology has the ability to play a significant role in safeguarding public health, improving safety measures, and aiding in the ongoing battle against COVID-19 and SARS in various community settings.

Morphology-based deep learning approach for predicting adipogenic and osteogenic differentiation of human mesenchymal stem cells (hMSCs)

Human mesenchymal stem cells (hMSCs) are multipotent progenitor cells with the potential to differentiate into various cell types, including osteoblasts, chondrocytes, and adipocytes. These cells have been extensively employed in the field of cell-based therapies and regenerative medicine due to their inherent attributes of self-renewal and multipotency. Traditional approaches for assessing hMSCs differentiation capacity have relied heavily on labor-intensive techniques, such as RT-PCR, immunostaining, and Western blot, to identify specific biomarkers. However, these methods are not only time-consuming and economically demanding, but also require the fixation of cells, resulting in the loss of temporal data. Consequently, there is an emerging need for a more efficient and precise approach to predict hMSCs differentiation in live cells, particularly for osteogenic and adipogenic differentiation. In response to this need, we developed innovative approaches that combine live-cell imaging with cutting-edge deep learning techniques, specifically employing a convolutional neural network (CNN) to meticulously classify osteogenic and adipogenic differentiation. Specifically, four notable pre-trained CNN models, VGG 19, Inception V3, ResNet 18, and ResNet 50, were developed and tested for identifying adipogenic and osteogenic differentiated cells based on cell morphology changes. We rigorously evaluated the performance of these four models concerning binary and multi-class classification of differentiated cells at various time intervals, focusing on pivotal metrics such as accuracy, the area under the receiver operating characteristic curve (AUC), sensitivity, precision, and F1-score. Among these four different models, ResNet 50 has proven to be the most effective choice with the highest accuracy (0.9572 for binary, 0.9474 for multi-class) and AUC (0.9958 for binary, 0.9836 for multi-class) in both multi-class and binary classification tasks. Although VGG 19 matched the accuracy of ResNet 50 in both tasks, ResNet 50 consistently outperformed it in terms of AUC, underscoring its superior effectiveness in identifying differentiated cells. Overall, our study demonstrated the capability to use a CNN approach to predict stem cell fate based on morphology changes, which will potentially provide insights for the application of cell-based therapy and advance our understanding of regenerative medicine.

An Ensemble of Deep Convolutional Neural Networks Models for Facial Beauty Prediction

Facial beauty prediction is an emerging topic. The pursuit of facial beauty is the nature of human beings. As the demand for aesthetic surgery has increased significantly over the past few years, an understanding beauty is becoming increasingly important in medical settings. This work proposes a new ensemble based on the pre-trained convolutional neural network (CNN) models to identify scores for facial beauty prediction. These ensembles were originally built from the following previously trained models: DenseNet-201, Inception-v3, MobileNetV2, and EfficientNetB7. According to the SCUT-FBP5500 benchmark dataset, the proposed model obtains a Pearson coefficient of 0.9469. This reveals that the suggested EN-CNNs model can be successfully applied in a variety of face-to-face applications.

Butterflies Recognition using Enhanced Transfer Learning and Data Augmentation

Butterflies’ recognition serves a crucial role as an environmental indicator and a key factor in plant pollination. The automation of this recognition process, facilitated by Convolutional Neural Networks (CNNs), can expedite this task. Several pre-trained CNN models, such as VGG, ResNet, and Inception, have been widely used for this purpose. However, the scope of previous research has been somewhat constrained, focusing only on a maximum of 15 classes. This study proposes to modify the CNN InceptionV3 model and combine it with three data augmentations to recognize up to 100 butterfly species. To curb overfitting, this study employs a series of data augmentation techniques. In parallel, we refine the InceptionV3 model by reducing the number of layers and integrating four new layers. The test results demonstrate that our proposed model achieves an impressive accuracy of 99.43% for 15 classes with only 10 epochs, exceeding prior models by approximately 5%. When extended to 100 classes, the model maintains a high accuracy rate of 98.49% with 50 epochs. The proposed model surpasses the performance of standard pre-trained models, including VGG16, ResNet50, and InceptionV3, illustrating its potential for broader application.

Cervical Cancer Classification From Pap Smear Images Using Deep Convolutional Neural Network Models

As one of the most common female cancers, cervical cancer often develops years after a prolonged and reversible pre-cancerous stage. Traditional classification algorithms used for detection of cervical cancer often require cell segmentation and feature extraction techniques, while convolutional neural network (CNN) models demand a large dataset to mitigate over-fitting and poor generalization problems. To this end, this study aims to develop deep learning models for automated cervical cancer detection that do not rely on segmentation methods or custom features. Due to limited data availability, transfer learning was employed with pre-trained CNN models to directly operate on Pap smear images for a seven-class classification task. Thorough evaluation and comparison of 13 pre-trained deep CNN models were performed using the publicly available Herlev dataset and the Keras package in Google Collaboratory. In terms of accuracy and performance, DenseNet-201 is the best-performing model. The pre-trained CNN models studied in this paper produced good experimental results and required little computing time.Graphical

Toward a Vision-Based Intelligent System: A Stacked Encoded Deep Learning Framework for Sign Language Recognition.

Sign language recognition, an essential interface between the hearing and deaf-mute communities, faces challenges with high false positive rates and computational costs, even with the use of advanced deep learning techniques. Our proposed solution is a stacked encoded model, combining artificial intelligence (AI) with the Internet of Things (IoT), which refines feature extraction and classification to overcome these challenges. We leverage a lightweight backbone model for preliminary feature extraction and use stacked autoencoders to further refine these features. Our approach harnesses the scalability of big data, showing notable improvement in accuracy, precision, recall, F1-score, and complexity analysis. Our model's effectiveness is demonstrated through testing on the ArSL2018 benchmark dataset, showcasing superior performance compared to state-of-the-art approaches. Additional validation through an ablation study with pre-trained convolutional neural network (CNN) models affirms our model's efficacy across all evaluation metrics. Our work paves the way for the sustainable development of high-performing, IoT-based sign-language-recognition applications.