Image Datasets Research Articles

Coffee bean graded based on deep net models

Coffee is a widely consumed beverage, and sorting coffee beans is a critical process that ensures high-quality graded coffee products. Coffee beans were graded into nine grades in robusta types. To automate the grading process, a deep learning-based approach was developed using a large dataset of high-resolution images and data augmentation techniques. In contrast to previous studies focusing on robusta type graded into six coffee bean grads, our research extends this framework by employing robusta type into nine grades with an outperformed accuracy. The proposed work uses four deep learning models, namely residual network 34(Resnet34), inception version 3 (Inception v3), efficient network bayesian optimization (EfficientNet-B0), and visual geometry group-16(VGG-16), where trained and evaluated for coffee bean classification into nine grades. The EfficientNet-B0 model exhibited outperformed accuracy, achieving 100% in distinguishing good and bad coffee beans, even in challenging lighting and background conditions.

TQU-HG dataset and comparative study for hand gesture recognition of RGB-based images using deep learning

Hand gesture recognition has great applications in human-computer interaction (HCI), human-robot interaction (HRI), and supporting the deaf and mute. To build a hand gesture recognition model using deep learning (DL) with high results then needs to be trained on many data and in many different conditions and contexts. In this paper, we publish the TQU-HG dataset of large RGB images with low resolution (640×480) pixels, low light conditions, and fast speed (16 fps). TQU-HG dataset includes 60,000 images collected from 20 people (10 male, 10 female) with 15 gestures of both left and right hands. A comparative study with two branches: i) based on Mediapipe TML and ii) Based on convolutional neural networks (CNNs) (you only look once (YOLO); YOLOv5, YOLOv6, YOLOv7, YOLOv8, YOLO-Nas, single shot multiBox detector (SSD) VGG16, residual network (ResNet)18, ResNext50, ResNet152, ResNext50, MobileNet V3 small, and MobileNet V3 large), the architecture and operation of CNNs models are also introduced in detail. We especially fine-tune the model and evaluate it on TQU-HG and HaGRID datasets. The quantitative results of the training and testing are presented (F1-score of YOLOv8, YOLO-Nas, MobileNet V3 small, ResNet50 is 98.99%, 98.98%, 99.27%, 99.36%, respectively on the TQU-HG dataset and is 99.21%, 99.37%, 99.36%, 86.4%, 98.3%, respectively on the HaGRID dataset). The computation time of YOLOv8 is 6.19 fps on the CPU and 18.28 fps on the GPU.

Weld Defect Downscaling Recognition for Phased Array Ultrasonic Data Based on Semantic Segmentation

Phased array ultrasonic testing (PAUT) is one of the most advanced ultrasonic inspection methods, providing inspection data in image sequences with increased resolution and coverage, consequently enhancing inspection efficiency. Due to the inherently abstract nature of ultrasound images, the analysis of PAUT weld data continues to rely on the expertise of experienced inspectors, particularly in the context of shipbuilding welds, which exhibit characteristics such as arbitrary lengths, varying thicknesses, and various weld methods, particularly when associated with hybrid laser-arc welding involving higher caps and narrower fill widths. This paper applies a semantic segmentation architecture to PAUT data for hybrid laser-arc welding to segment defects automatically. A downscaling PAUT image dataset incorporating PAUT imaging principles with scan processes is established. The proposed architecture adopts an encoder-decoder architecture and segments defects downscaling, enabling direct localization and sizing defects in long PAUT data. Meanwhile, it handles PAUT image sequences of arbitrary length, eliminating the need for horizontal scale normalization while preserving the integrity of defect shapes. The proposed architecture was validated in the dataset to demonstrate the superiority of the proposed method in locating and sizing defects in PAUT weld data, offering a promising automated solution to a challenge in ultrasonic inspection methodologies.

A new deep learning model with interface for fine needle aspiration cytology image-based breast cancer detection

Cytological evaluation through microscopic image analysis of fine needle aspiration cytology (FNAC) is pivotal in the initial screening of breast cancer. The sensitivity of FNAC as a screening tool relies on both image quality and the pathologist’s expertise. To enhance diagnostic accuracy and alleviate the pathologist’s workload, a computer-aided diagnosis (CAD) system was developed. A comparative study was conducted, assessing twelve candidate pre-trained models. Utilizing a locally gathered FNAC image dataset, three superior models-MobileNet-V2, DenseNet-121, and Inception-V3-were selected based on their training, validation, and testing accuracies. Further, these models underwent evaluation in four transfer learning scenarios to enhance testing accuracy. While the outcomes were promising, they left room for improvement, motivating us to create a novel deep convolutional neural network (CNN). The newly proposed model exhibited robust performance with testing accuracy at 85%. Our research concludes that the most lightweight, high-accuracy model is the one we propose. We’ve integrated it into our user-friendly Android App, “Breast Cancer Detection System,” in TensorFlow Lite format, with cloud database support, showcasing its effectiveness. Implementing an artificial intelligent (AI)-based diagnosis system with a user-friendly interface holds the potential to enhance early breast cancer detection using FNAC.

CellSNAP: a fast, accurate algorithm for 3D cell segmentation in quantitative phase imaging.

Three-dimensional quantitative phase imaging (QPI) has rapidly emerged as a complementary tool to fluorescence imaging, as it provides an objective measure of cell morphology and dynamics, free of variability due to contrast agents. It has opened up new directions of investigation by providing systematic and correlative analysis of various cellular parameters without limitations of photobleaching and phototoxicity. While current QPI systems allow the rapid acquisition of tomographic images, the pipeline to analyze these raw three-dimensional (3D) tomograms is not well-developed. We focus on a critical, yet often underappreciated, step of the analysis pipeline that of 3D cell segmentation from the acquired tomograms. We report the CellSNAP (Cell Segmentation via Novel Algorithm for Phase Imaging) algorithm for the 3D segmentation of QPI images. The cell segmentation algorithm mimics the gemstone extraction process, initiating with a coarse 3D extrusion from a two-dimensional (2D) segmented mask to outline the cell structure. A 2D image is generated, and a segmentation algorithm identifies the boundary in the plane. Leveraging cell continuity in consecutive -stacks, a refined 3D segmentation, akin to fine chiseling in gemstone carving, completes the process. The CellSNAP algorithm outstrips the current gold standard in terms of speed, robustness, and implementation, achieving cell segmentation under 2s per cell on a single-core processor. The implementation of CellSNAP can easily be parallelized on a multi-core system for further speed improvements. For the cases where segmentation is possible with the existing standard method, our algorithm displays an average difference of 5% for dry mass and 8% for volume measurements. We also show that CellSNAP can handle challenging image datasets where cells are clumped and marred by interferogram drifts, which pose major difficulties for all QPI-focused AI-based segmentation tools. Our proposed method is less memory intensive and significantly faster than existing methods. The method can be easily implemented on a student laptop. Since the approach is rule-based, there is no need to collect a lot of imaging data and manually annotate them to perform machine learning based training of the model. We envision our work will lead to broader adoption of QPI imaging for high-throughput analysis, which has, in part, been stymied by a lack of suitable image segmentation tools.

Multi-granularity tooth analysis via YOLO-based object detection models for effective tooth detection and classification

Accurate detection and classification of teeth is the first step in dental disease diagnosis. However, the same class of tooth exhibits significant variations in surface appearance. Moreover, the complex geometrical structure poses challenges in learning discriminative features among the different tooth classes. Due to these complex features, tooth classification is one of the challenging research domains in deep learning. To address the aforementioned issues, the presented study proposes discriminative local feature extraction at different granular levels using YOLO models. However, this necessitates a granular intra-oral image dataset. To facilitate this requirement, a dataset at three granular levels (two, four, and seven teeth classes) is developed. YOLOv5, YOLOv6, and YOLOv7 models were trained using 2,790 images. The results indicate superior performance of YOLOv6 for two-class classification problems. The model generated a mean average precision (mAP) value of 94%. However, as the granularity level is increased, the performance of YOLO models decreases. For, four and seven-class classification problems, the highest mAP value of 87% and 79% was achieved by YOLOv5 respectively. The results indicate that different levels of granularity play an important role in tooth detection and classification. The YOLO’s performance gradually decreased as the granularity decreased especially at the finest granular level.

Deep Learning Algorithms for Skin Disease Classification

Skin diseases are a serious concern of public health worldwide, and successful treatment needs a correct and timely diagnosis. Traditional diagnostic methods mostly depend on dermatologist’s visual observation and this leads to subjective interpretations coupled with time-consuming processes. Deep learning algorithms have lately been known as powerful means for automated medical image analysis that present more accurate and quicker results at the same time. This study analyses the usage of state-of-the-art deep learning algorithms like YOLOv8, Deep CNN, and ResNet50 used for classification of skin diseases using dermatological images. Classifying the skin conditions relies heavily on the ability to identify and extract essential features. Different skin conditions were covered under large dataset thus providing a comprehensive foundation for training and validation aimed at ensuring that the models could generalize well across different diseases. Each algorithm also employs transfer learning techniques by utilizing pre-trained models based on large image datasets in order to improve adaptability and generalization over new data types. The use of deep learning algorithms in classifying skin diseases represents a significant method to achieve efficient and accurate diagnosis with benefits to both patients and healthcare professionals as is the trend in medical image analysis. The advanced deep learning models introduced in this paper excel at classifying complex skin diseases, outperforming the machine learning approaches in performance.

Deep Learning Enabled Diabetic Retinopathy Diagnosis Using Retinal Fundus Images

Abstract: This paper pivots on (DR) is a leading cause of blindness worldwide. Early detection and effective management are crucial in preventing damage to the retina. In recent years, (DL) techniques have shown promise in aiding the DR. This paper aims to investigate the diagnosis of using (CNNs) for automated DR screening. The study utilized a available patient dataset of retinal images to train and testing CNN model for DR detection and classification. Results showed that given (DL) techniques achieved with accurately in both tasks, outperforming existing state-of-the-art methods. The findings suggest the (DL) algorithms can aid in the early detection and management of DR, potentially reducing the burden of the condition on healthcare systems and improving patient outcomes

Micronutrient Deficiency Using Image Processing and Neural Network

Abstract: Vitamin and micronutrient deficiencies are a significant global health concern, leading to various adverse health consequences. Early detection and intervention are essential in addressing these issues. The project introduces an intelligent system that utilizes advanced deep learning techniques to identify and differentiate vitamin deficiencies in human tissue through image analysis. The approach involves an initial step of image clustering to separate and isolate problem areas from the input images. The goal is to evaluate the productivity ofimage segmentation methods, extract relevant characteristics, and compare classification results with other methods. To accomplish the objectives, a diverse dataset of facial images is gathered and preprocessed, focusing on individuals both with and without visible signs of vitamin deficiencies Following this, a CNN algorithm, inspired by models like Alex Net, is created and trained using the pre-processed dataset. The CNN isused to identify and classify features based on different types of vitamin deficiencies, enabling an automated andaccurate assessment based on facial images

Classification of Dermoscopic Images Using CNN-SVM

Traditional machine learning methods like GLCM and ABCD rules have long been employed for image classification tasks. However, they come with inherent limitations, primarily the need for manual feature extraction. This manual feature extraction process is time-consuming and relies on expert domain knowledge, making it challenging for non-experts to use effectively. Deep learning methods, specifically Convolutional Neural Networks (CNN), have revolutionized image classification by automating the feature extraction. CNNs can learn hierarchical features directly from the raw pixel values, eliminating the need for manual feature engineering. Despite their powerful capabilities, CNNs have limitations, mainly when working with small image datasets. They may overfit the data or struggle to generalize effectively. In light of these considerations, this study adopts a hybrid approach that leverages the strengths of both deep learning and traditional machine learning. CNNs are automatic feature extractors, allowing the model to capture meaningful image patterns. These extracted features are then fed into a Support Vector Machine (SVM) classifier, known for its efficiency and effectiveness in handling small datasets. The results of this study are encouraging, with an accuracy of 0.94 and an AUC score of 0.94. Notably, these metrics outperform Abbas' previous research by a significant margin, underscoring the effectiveness of the hybrid CNN-SVM approach. This research reinforces that SVM classifiers are well-suited for tasks involving limited image data, yielding improved classification accuracy and highlighting the potential for broader applications in image analysis.

Automating Helmet Usage Detection: A YOLOv8 Based Framework

Abstract: The biggest threat to Powered two-wheeler(PTW) riders is a head injury. With the humongous amount of daily commuters on such PTWs, the risk only increases substantially. In such cases, helmets reduce the risk of head injuries in accidents but not every citizen is keen on their use. In India, where the use of motorcycles is widespread, ensuring helmetuse remains a challenge. This study explores the effectiveness of YOLOv8(You Only Look Once) a state-of-the-art single-stage object detection algorithm for helmet detection on Indian roads [14]. The CNN(Convolutional Neural Networks) based technologyimplements many other efficient, reliable and quick algorithmsto train, validate and predict the object’s detection, segmentationand classification task. We leverage the use of a publicly available Indian helmet dataset from Kaggle, containing 942 images with annotations. This an image dataset of real-life powered two-wheeler riders. The dataset has images and video frames for training on 5 different objects/classes along with annotations that specify thoseobjects. This multi-class approach offers valuable insight intohelmet usage patterns paves shows us the way for real-world applications for automated traffic monitoring. Real-time helmet and number plate detection can become a game changer in trafficmonitoring and can strengthen basic safety laws

Space-time image velocimetry in blurred scenes based on BSTI-DCGAN data augmentation

Due to the limited sample quantity and the complex data collection process of the blurred space-time image (BSTI) dataset, the deep learning-based space-time image velocimetry (STIV) results in larger errors when applied to blurry videos. To enhance the measurement accuracy, we propose the use of STIV in blurred scenes based on BSTI-deep convolutional generative adversarial network (DCGAN) data augmentation. Firstly, BSTI-DCGAN is developed based on the DCGAN. This network utilizes a bilinear interpolation-convolution module for upsampling and integrates coordinated attention and multi-concatenation attention to enhance the resemblance between generated and real images. Next, further expanding the dataset by using artificially synthesized space-time images subsequently, all space-time images are transformed into spectrograms to create a training dataset for the classification network. Finally, the primary spectral direction is detected using the classification network. The experimental results indicate that our approach effectively augments the dataset and improves the accuracy of practical measurements. Under the condition of video blur, the relative errors of the average flow velocity and discharge are 3.92% and 2.72%, respectively.

Enhancing OCR with line segmentation mask for container text recognition in container terminal

Optical Character Recognition (OCR) plays a pivotal role in enhancing the operational efficiency of container ports. However, challenges such as angle limitations and the complexity of container fonts in traditional OCR systems lead to tilted text and text adhesion, thereby reducing the recognition rate. Recognizing containers at a high speed is equally crucial for port operations. In this study, we address these challenges by introducing an Enhanced OCR (EOCR) system, incorporating Line Segmentation Mask (LSM)-based detection and Scanline-based recognition. LSM tackles the issue of text adhesion caused by traditional segmentation, while recognition based on scan lines accelerates efficiency. Additionally, we propose the arbitrary angle quadrilateral fitting algorithm targeting sloping quad areas in images taken at a container terminal. Experimental results on a dataset of container images from the Shanghai Port demonstrate superior performance compared to existing algorithms, achieving a recognition accuracy rate of up to 98.7%. Furthermore, an ablation study confirms that our EOCR significantly enhances recognition accuracy while ensuring real-time performance.

Integration of feature enhancement technique in Google inception network for breast cancer detection and classification

Breast cancer is a major public health concern, and early detection and classification are essential for improving patient outcomes. However, breast tumors can be difficult to distinguish from benign tumors, leading to high false positive rates in screening. The reason is that both benign and malignant tumors have no consistent shape, are found at the same position, have variable sizes, and have high correlations. The ambiguity of the correlation challenges the computer-aided system, and the inconsistency of morphology challenges an expert in identifying and classifying what is positive and what is negative. Due to this, most of the time, breast cancer screen is prone to false positive rates. This research paper presents the introduction of a feature enhancement method into the Google inception network for breast cancer detection and classification. The proposed model preserves both local and global information, which is important for addressing the variability of breast tumor morphology and their complex correlations. A locally preserving projection transformation function is introduced to retain local information that might be lost in the intermediate output of the inception model. Additionally, transfer learning is used to improve the performance of the proposed model on limited datasets. The proposed model is evaluated on a dataset of ultrasound images and achieves an accuracy of 99.81%, recall of 96.48%, and sensitivity of 93.0%. These results demonstrate the effectiveness of the proposed method for breast cancer detection and classification.

Efficient Multi-scale Network for Semantic Segmentation of fine-Resolution Remotely Sensed Images

Abstract Semantic segmentation of remote sensing urban scene images has diverse practical applications, including land cover mapping, urban change detection, environmental protection, and economic evaluation. However, classical semantic segmentation networks encounter challenges such as inadequate utilization of multi-scale semantic information and imprecise edge target segmentation in high-resolution remote sensing images. In response, this article introduces an Efficient Multi-scale Network (EMNet) tailored for semantic segmentation of common features in remote sensing images.To address these challenges, EMNet integrates several key components. Firstly, the Efficient Atrous Spatial Pyramid Pooling module(EASPP) is employed to enhance the relevance of multi-scale targets, facilitating improved extraction and processing of context information across different scales. Secondly, the Efficient Multi-Scale Attention mechanism (EMA)and multi-scale jump connections are utilized to fuse semantic features from various levels, thereby achieving precise segmentation boundaries and accurate position information. Finally, an encoder-decoder structure is incorporated to refine the segmentation results.The effectiveness of the proposed network is validated through experiments conducted on the publicly available DroneDeploy image dataset. Results indicate that EMNet achieves impressive performance metrics, with Mean Intersection over Union (MIoU), Mean Pixel Accuracy (MPA), Mean Precision (MPrecision), and Mean Recall (MRecall) reaching 75.99%, 85.08%, 86.76%, and 85.07%, respectively. Comparative analysis demonstrates that the network proposed in this article outperforms current mainstream semantic segmentation networks on the DroneDeploy dataset.

Precancerous Change Detection Technique on Mammography Breast Cancer Images based on Mean Ratio and Log Ratio using Fuzzy c Mean Classification with Gabor Filter.

The growing rate of breast cancer necessitates immediate global attention. Mammography images are used to determine the stage of malignancy. Breast cancer stages must be identified in order to save a person's life. This article's main goal is to identify different techniques to obtain the difference between two breast cancer mammography images taken of the same individual at different times. This is the first effort to identify breast cancer in mammography images using change detection techniques. The Mammogram Image Change Detection (ICD) technique is also a recent advancement to prevent breast cancer in the early stage and precancerous level in medical images. The main purpose of this work is to observe the changes between breast cancer images in different screening periods using different techniques. Mammogram Breast Cancer Image Change Detection (MBCICD) methods usually start with a Difference Image (DI) and classify the pixels in the DI into changed and unaffected classes using unsupervised fuzzy c means (FCM) clustering methods based on texture features taken from the log and mean ratio difference pictures. Two operators, mean ratio and log ratio, were used to check the changes in the images. The Gabor wavelet is utilized as a feature extraction technique among several standards. Using the Gabor wavelet ratio operators is a useful method for altering the detection of breast cancer in mammography images. Currently, it is challenging to obtain real malignant images of the same person for testing or training. In this study, two images are utilized. To clearly see the changes, one is an image from the MIAS breast cancer mammography images dataset, and the other is a self-generated change image. The research aims to examine the image results and other quantitative analysis results of proposed change detection methods on cancer images. The Mean Ratio Accuracy result is 0.9738, and the Log ratio PCC is 0.9737. The classification results are the Log Ratio + Gabor Filter + FCM is 0.9737, and Mean Ratio +Gabor Filter + FCM is 0.9719. The mean Ratio Accuracy result is 0.9738, Log ratio is 0.9737. Log Ratio + Gabor Filter + FCM is 0.9737, Mean Ratio +Gabor Filter + FCM is 0.9719. Comparing the PCC of proposed change detection methods with the FDA-RMG method on the same dataset, the accuracy is 0.9481 only. The study concludes that variations in mammography breast cancer images could be successfully identified using the ratio operators with Gabor wavelet features.

Supersonic combustion field evolution prediction in scramjet engine using a deblurring multi-scale attention network

Traditional deep learning technologies often overlook crucial spatiotemporal information when reconstructing supersonic combustion flow fields. These models can only reconstruct the current flow field, resulting in poor visual quality due to the atomization effect. This research introduces experimental studies on scramjet combustion flows with variable injection pressures at the inlet of a scramjet Mach 2.5, creating image datasets of flame evolution processes over different predicted time spans (5, 10, and 15 ms). It develops a multi-scale attention algorithm with deblurring (DB-MSAN) capabilities to predict the flame image after a certain period using the wall pressure signal from the previous moment. The algorithm employs two multi-head attention channels at different scales to extract timing information, encouraging the model to focus on the anisotropy of the input data and to identify more significant features. The DB module utilizes the adapted atmospheric scattering model to jointly estimate the transmission map and atmospheric light, achieving an exceptional defogging effect with a minimal computational expense. The model’s overall performance is assessed regarding efficiency and accuracy, comparing DB-MSAN’s performance differences with those of Chen’s and the ResNet16 models across multi-span, large-span, various test sets, and sparse wall measurement points. Experimental results indicated that, compared to other methods, DB-MSAN achieves a maximum peak signal-to-noise index improvement of 76.10 % and an SSIM index improvement of 155.26 % when predicting average-quality images. The prediction accuracy remains stable, even with sparse pressure input. In addition, DB-MSAN strikes an optimal balance between the number of parameters and model accuracy; it represents only 16.97 % of the volume compared to Chen’s model, which is 585 MB.

Proximal femur fracture detection on plain radiography via feature pyramid networks.

Hip fractures exceed 250,000 cases annually in the United States, with the worldwide incidence projected to increase by 240-310% by 2050. Hip fractures are predominantly diagnosed by radiologist review of radiographs. In this study, we developed a deep learning model by extending the VarifocalNet Feature Pyramid Network (FPN) for detection and localization of proximal femur fractures from plain radiography with clinically relevant metrics. We used a dataset of 823 hip radiographs of 150 subjects with proximal femur fractures and 362 controls to develop and evaluate the deep learning model. Our model attained 0.94 specificity and 0.95 sensitivity in fracture detection over the diverse imaging dataset. We compared the performance of our model against five benchmark FPN models, demonstrating 6-14% sensitivity and 1-9% accuracy improvement. In addition, we demonstrated that our model outperforms a state-of-the-art transformer model based on DINO network by 17% sensitivity and 5% accuracy, while taking half the time on average to process a radiograph. The developed model can aid radiologists and support on-premise integration with hospital cloud services to enable automatic, opportunistic screening for hip fractures.

A fiber recognition framework based on multi-head attention mechanism

Convolutional neural networks have been extensively studied in textile fiber recognition. However, the down-sampling performed by the convolutional and pooling layers on the extracted features result in a significant loss of fine-grained fiber features. To address this issue, a multi-head attention framework for fiber recognition has been proposed. First, textile surface images are captured using optical magnifiers and smart devices such as smartphones. Second, fiber features and label embeddings are learned separately by multi-head self-attention. Finally, the label embeddings are used to query the presence of fiber types, and pool type-related features in the multi-head cross-attention module to classify fibers. The experimental results demonstrate that the proposed method performs exceptionally well on the textile surface image dataset, with a mean average precision accuracy improvement of 6% compared with the convolutional neural network-based fiber recognition method Cu-Net, and a 5% improvement compared with the fiber recognition method FiberCT combining convolutional neural networks and attention mechanisms. It is worth mentioning that the fiber images captured at 200× magnification in the collected dataset are most favorable for models to recognize fiber. The proposed method achieves a mean average precision fiber recognition accuracy of 80.2% on images at 200× magnification.

A Fresh Look at Tomato Leaf Disease Recognition using Vision Transformers

Tomatoes is one of the major economically significant vegetables produced worldwide, contributing greatly to increased agricultural production and food security. However, tomato plants are unfortunately prone to a number of diseases, including several that target the leaves, which can significantly reduce crop productivity and quality. Recently, deep learning techniques have revolutionized the fields of computer vision and image analysis. By automatically learning hierarchical representations from raw pixel data. Transformer is a new deep learning technique that opens new possibilities for image understanding using self-attention mechanisms to capture global dependencies within input sequences. This approach is exemplified by the Vision Transformer (ViT). In this study, we utilize and evaluate the effectiveness of six variations of the Vision Transformer (ViT) architecture in the task of tomato leaf disease recognition. The variants include Mobile ViT, EANet, Swin ViT, ViT, Shift ViT, and Compact ViT. Utilizing a publicly available, multiple-source dataset of tomato leaf images containing various disease patterns. Performance for all models was evaluated and compared in classifying various types of tomato leaf diseases in terms of accuracy, loss, precision, recall, and F1-Score, and the results showed that. The CompactVit has achieved the best accuracy of 97% and precision of 97% and 96% for recall. While the mobile ViT has the lowest performance among all variations in tomato disease recognition, overall, ViT is showing its promise, and it can be utilized on a large scale for smart agriculture, which opens the door for further exploration of this area.