Fused deep neural networks for sustainable and computational management of heat-transfer pipeline diagnosis

In the development of social economy and scientific and technological innovation, the image processing mode and classification model chosen by network technology platform is becoming more and more changeable, but in essence, it is necessary to obtain characteristic information in effective image recognition and choose high-quality network algorithm and processing technology to complete image processing and image classification. Therefore, on the basis of understanding the current research trend of computer image processing and image classification model methods, this paper conducts in-depth discussion on the image processing methods and image classification model training design with artificial intelligence as the core and takes the image classification model of transfer learning as an example for practical exploration. The final results show that the image processing method and image classification model based on artificial intelligence have strong performance advantages in practical application.KeywordsArtificial intelligenceImage processingImage classificationThe migration study

Despite widespread use of standardized classification systems, risk stratification of thyroid nodules is nuanced and often requires diagnostic surgery. Genomic sequencing is available for this dilemma however, costs and access restricts global applicability. Artificial intelligence (AI) has the potential to overcome this issue nevertheless, the need for black-box interpretability is pertinent. We aimed to create an ultrasonographic segmentation and classification model that offers explainability and risk accountability. Four hundred and fourteen ultrasonography images were collected from 105 patients undergoing thyroidectomy, divided into training and testing groups. Classification ground truth used is exclusively surgical histopathology. Relevant nodules were manually annotated by a dedicated study radiologist and surgeon. Three AI architectures with and without block attention modules were trained to identify the relevant nodule and the best performing was selected for the subsequent task in classifying identified nodules into benign or malignant. Gradient-Weighted Class Activation Map is used to provide saliency mapping for visual interpretability. Superior performance was recorded by the block attention model which stratified thyroid nodules into benign versus malignant with an accuracy of 93% versus 90%, F-score 90% versus 89%, sensitivity 93% versus 91% and specificity 92% versus 91% on a training dataset versus a testing dataset respectively. Visual interpretability maps demonstrate salient areas for a benign nodule diagnosis overlaps spongiform areas and malignant diagnosis salient areas overlap solid components of a partially cystic-solid nodule and microcalcifications within nodules. These findings are consistent with established diagnostic criteria for benign and malignant nodules. We developed an image segmentation and classification model for the risk stratification of thyroid nodules benchmarking surgical histopathology as ground truth and providing visual interpretability.

Performance of deep learning models for classifying and detecting common weeds in corn and soybean production systems

Cancer is one of humanity’s biggest challenges. Although there are many ways to prevent cancer, some types of cancer are still largely incurable. One of the most common types of cancer is breast cancer and the most important action toward a favorable outcome is early diagnosis. Correct diagnosis is one of the most critical processes in breast cancer treatment. There are many studies in the literature to predict the type of breast tumors. In this study, Artificial Intelligence and machine learning were used to increase the incidences of correct diagnosis. A convolutional neural network (CNN) was used in the models to classify and segment breast ultrasound images. Two separate models were created; one for image classification, and another for image segmentation, using TensorFlow. The models were based on current Convolutional Neural Network models published previously. The image segmentation model was based on a customized U-Net architecture. Attention gates were added to this customized U-Net model. The image classification model was constructed using a 50-layer ResNet model. The resultant CNN yielded an accuracy of 99.35% for image segmentation. For image classification, the separately constructed CNN model yielded an accuracy of 97.43%. Hence, it was possible to obtain high accuracies to classify and segment breast cancer from ultrasound images using our CNN models.

Image Classification (IC) is most prominent among other Artificial Intelligence (AI) domains. Mainly, IC participates rigorously for the development of society in a variety of application areas such as finance, marketing, health, industrial automation, education, and safety and security. Typically, an IC model takes image input data and tunes itself as per the required application task and classify accordingly. Among the various categories of images, color image category is better due to the capability of capturing more details, which are essential for classification purpose. However, the modern world demands Realtime or online image classification, which involves Imagery Streams. The highly likely uncertainty in Imagery Streams is due to non-stationary environment, for example, certain features or class boundaries which are valid at one-time step are not adequate for another time step. These uncertainties in Imagery Streams have deleterious effects on IC models, which causes performance degradation in terms of accuracy or make IC models, not in further use. Therefore, to overcome these issues, IC models need to adapt to changes caused by uncertainties in Imagery Streams. This paper focuses on the understanding the possible scenarios of such uncertainties in Color Imagery Streams, investigates the deleterious effects due to changes in Color Imagery Streams and provides the possible mitigation approach to overcome the issues in IC models. The contribution of this research is the first step towards an adaptive model development to mitigate the deleterious effects of uncertainty in Color Imagery Streams. This model will benefit many application areas and will directly contribute to the daily life of a society.

Recently, the bag-of-visual words (BoW) models have widely been studied in computer vision area. Owing to the limit of the BoW models that only consider the distributions of visual words in images, the Latent Dirichlet Allocation (LDA) model has drawn an attention to discover the structure of the visual word distributions over latent topics which can represent semantic objects in images. In order to reflect the spatial information of images, the LDA model has been extended to so-called a spatial LDA model for image segmentation, which is not applicable for image classification. Therefore, in this paper, we propose a spatial class LDA (scLDA) model for image classification where the topic distributions over visual words are found per image segments and a class-specific-simplex LDA (cssLDA) model is applied for image classification. From our experimental results, the proposed scLDA model outperforms the previous LDA models in terms of correct classification rates.

In recent years, generative adversarial networks (GANs) have gained tremendous popularity for various imaging related tasks such as artificial image generation to support AI training. GANs are especially useful for medical imaging-related tasks where training datasets are usually limited in size and heavily imbalanced against the diseased class. We present a systematic review, following the PRISMA guidelines, of recent GAN architectures used for medical image analysis to help the readers in making an informed decision before employing GANs in developing medical image classification and segmentation models. We have extracted 54 papers that highlight the capabilities and application of GANs in medical imaging from January 2015 to August 2020 and inclusion criteria for meta-analysis. Our results show four main architectures of GAN that are used for segmentation or classification in medical imaging. We provide a comprehensive overview of recent trends in the application of GANs in clinical diagnosis through medical image segmentation and classification and ultimately share experiences for task-based GAN implementations.

Manual object identification labelling is laborious, time-consuming and prone to inconsistencies hindering advancements in various computer vision tasks.These inconsistencies can lead to inaccurate models with poor performance. Considering these potential consequences, highlights the importance of addressing labelling challenges for ethical and responsible AI development. To address this our study evaluates several popular platforms for their suitability in tackling these challenges. Roboflow, Makesense.ai, SentiSight.ai, Labelbox and SuperAnnotate are the five different data labelling platforms that have been taken for assessment. The study identifies strengths and weaknesses of each platform in the context of basketball detection using YOLO v8, a deep learning model for object detection, image classification, and image segmentation. Each platform is analysed based on features, ease of use, pricing, and support for image annotation, object detection, and YOLO v8 integration. After analysing these factors, a final recommendation is made, highlighting the platform that demonstrably offers the best balance of features, efficiency, and cost-effectiveness for this specific task. The study helps in deeper exploration of the potential of YOLO v8. It is mainly aimed at assisting the Video Assistant Referees(VARs) for accurate and unbiased decision-making and also empowers the development of AI technology across the domain of sports.

We introduce the Plant Disease Detection Platform (PDDP) that allows users to send photos of sick plant leaves or textual descriptions of their appearance to obtain information about an infection that hit the vegetation and treatment tips. The backend of the platform in terms of deep learning includes image classification and text similarity models. The image classification model has two parts: feature extractor and classifier. The feature extractor was trained using the triplet loss function along with transfer learning when the weights of the network are initialized from the MobileNetV2 pretrained on the ImageNet dataset. The classifier is a simple multilayer perceptron. The test on 100 random plant images from the Internet shows 98% of the classification accuracy. We did the post-training static quantization in order to reduce the overall model size and increase the inference performance. The final model has a size of 7 Mb and works 5 times faster than the initial model without significant loss of accuracy. The text similarity model is a BERT-based transformer for obtaining vector representation of input texts for further similarity calculation between user requests and disease descriptions on the PDDP.

Forest fires pose a significant threat to ecosystems, wildlife, property, and human lives worldwide. Leveraging advancements in artificial intelligence and machine learning, our research presents a comprehensive approach to forest fire detection and management. We employ a state-ofthe-art image classification model, YOLOv8, to swiftly identify fire occurrences within forest imagery, achieving high accuracy rates. Concurrently, we develop a predictive model using logistic regression to forecast the likelihood of fire outbreaks based on environmental factors. Integration of these technologies holds promise for proactive forest fire management. Future prospects include the integration of our YOLOv8 model with UAVs for real-time monitoring and early detection of fire occurrences, thus enhancing environmental conservation and safety measures.

Recently, number of medical X-ray images being generated is increasing rapidly due to the advancements in radiological equipment in medical centres. Medical X-ray image classification techniques are needed for effective decision making in the healthcare sector. Since the traditional image classification models are ineffective to accomplish maximum X-ray image classification performance, deep learning (DL) models have emerged. In this study, an Arithmetic Optimization Algorithm with Deep Learning-Based Medical X-Ray Image Classification (AOADL-MXIC) model has been developed. The proposed AOADL-MXIC model investigates the available X-ray images for the identification of diseases. Initially, the AOADL-MXIC model executes the pre-processing step using the Gabor filtering (GF) technique to eliminate the presence of noise. In the next level, the Capsule Network (CapsNet) model is utilized to derive feature vectors from the input X-ray images. Furthermore, for optimizing the hyperparameters related to the CapsNet approach, the AOA is exploited. Finally, the bidirectional gated recurrent unit (BiGRU) model is employed for the classification of medical X-ray images. The experimental result analysis of the AOADL-MXIC technique on a set of medical images stated the promising performance over the other models.KeywordsX-ray imagesArithmetic optimization algorithmDeep learningFeature extractionHyperparameter tuning

The article provides a method for assessing the intensity of apple blossom for the thinning technological operation, including dataset preparation and training of YOLOv8-seg convolutional neural network models (n, s, m, l, x) for image segmentation. Transfer learning technique was applied in the research, utilizing pretrained models on the COCO dataset (Common Objects in Context). The apple blossom image dataset was captured using a GoPro HERO 11 camera. Image annotation was performed on the Roboflow platform using tools for bounding box and polygon annotation and labeling. To expand the dataset and improve the models' generalization during training, augmentation of original images was conducted, including horizontal flipping, horizontal rotation by 90°, rotation from -15° to +15°, adding noise up to 5% of pixels, blurring up to 2.5 pixels, horizontal and vertical shifts from -10° to 10°, and color hue adjustment from -15° to +15°. Binary classification metrics such as Precision and Recall were used to evaluate the performance of trained models in recognizing apple blossoms in images using bounding boxes and mask segmentation. The Loss(Box/Mask) loss function was used to assess model errors in determining bounding boxes and segmentation masks of objects in images during training. The hyperparameters of the YOLOv8-seg model for image recognition, classification, and segmentation of apple blossom images were identified through the YOLOv8x-seg (mAP50 metric = 0.591) and YOLOv8l-seg (mAP50 metric = 0,584) models demonstrate higher performance in apple blossom recognition. The frame processing speed (Frame Rate, FR) of convolutional neural network models ranged from 10.27 (YOLOv8x-seg model) to 57.32 (YOLOv8n-seg model). The average absolute error of the models in recognizing apple blossoms and counting their quantity in the test dataset does not exceed 9 %.

Deep Learning (DL) models for semantic image segmentation are an emerging trend in response to the explosion of multi-class, high resolution image and video data. However, segmentation models are highly compute-intensive, and even the fastest Volta GPUs cannot train them in a reasonable time frame. In our experiments, we observed just 6.7 images/second on a single Volta GPU for training DeepLab-v3+ (DLv3+), a state-of-the-art Encoder-Decoder model for semantic image segmentation. For comparison, a Volta GPU can process 300 images/second for training ResNet-50, a state-of-the-art model for image classification. In this context, we see a clear opportunity to utilize supercomputers to speed up training of segmentation models. However, most published studies on the performance of novel DL models such as DLv3+ require the user to significantly change Horovod, MPI, and the DL model to improve performance. Our work proposes an alternative tuning method that achieves near-linear scaling without significant changes to Horovod, MPI, or the DL model. In this paper, we select DLv3+ as the candidate TensorFlow model and implement Horovod-based distributed training for DLv3+. We observed poor default scaling performance of DLv3+ on the Summit system at Oak Ridge National Laboratory. To address this, we conducted an in-depth performance tuning of various Horovod/MPI knobs to achieve better performance over the default parameters. We present a comprehensive scaling comparison for Horovod with MVAPICH2-GDR up to 132 GPUs on Summit. Our optimization approach achieves near-linear (92%) scaling with MVAPICH2-GDR. We achieved a “mIOU” accuracy of 80.8% for distributed training, which is on par with published accuracy for this model. Further, we demonstrate an improvement in scaling efficiency by 23.9% over default Horovod training, which translates to a 1.3× speedup in training performance.

Honey is a sweet, sticky, yellowish-brown fluid made by bees and other insects from nectar collected from flowers, which often been used for food supplement or natural drug. Each nectar flower produces different kind of honey, for each honey has particular benefits. In this study, a detection system of honey botanical origin was proposed based on the spectral transmittance profile using hyperspectral imaging and machine learning. An acquiring image system consists of the transmittance module, halogen lamp, object slider, and hyperspectral imaging system. The image was recorded in 448 bands with a wavelength range from 400 nm to 1000 nm. An image processing method performs image correction, segmentation, feature extraction, feature reduction, and classification model. A classification model used a Pattern Recognition Network with a single hidden layer. A Bayesian regularization backpropagation was conducted to train the model. Five-type of the honey botanical origin from three different brands was collected to evaluate the proposed system. Three samples were prepared and measured for each botanical from each brand to create the honey dataset. A confusion matrix was used to measure classification performance. Based on the experiment result, the accuracy of botanical origin classification is 94.1%. The result shows an excellent result for the classification system.

Detection of spinal cord injury (SCI) is one of the major problems in children and adults due to variation in shape and orientation. As the types of spinal cord injuriesare increasing, it is difficult to find and predict the new type of disorder due to high dimensionality and sparsity problems. Most of the existing models are used to extract either the limited number of features or over segmented features on the SCI data. These models are not applicable to filter the essential features space with less segmented regions for injury disorder prediction. In such a scenario, we propose a hybrid threshold-based image segmentation and classification model is implemented for disorder prediction. In this model, a hybrid Ostu's thresholding method and expectation maximization (EM) approach and robust decision tree classifier are used to filter the essential features for disorder prediction. A hybrid CNN framework is used to extract the feature sets on the segmented features. Finally, a probabilistic classification model is used to predict the disease severity on the segmented image features. Experimental results illustrate the efficiency of proposed disorder prediction model with the existing models with 0.97 accuracy and 0.98 precision rate on the SCI dataset.