Aiming to address the issue of low accuracy in existing algorithms due to the limited scale of specific apple disease datasets and complex background information, a deep learning method integrating apple disease class...
详细信息
ISBN:
(纸本)9798400709234
Aiming to address the issue of low accuracy in existing algorithms due to the limited scale of specific apple disease datasets and complex background information, a deep learning method integrating apple disease classification and segmentation is proposed. Initially, the method employs the CycleGAN network for data enhancement to combat overfitting problems in deep learning. Secondly, it constructs a deep residual disease classification network embedded in the convolutional block attention module. In essence, the ResNet50(*) module extracts fine-grained visual features and incorporates both the channel attention mechanism and spatial attention mechanism to focus on two critical features, thereby enhancing the disease classification performance of the model across various scenarios. Lastly, a hierarchical semantic segmentation model is introduced to achieve a more accurate segmentation effect through segmented segmentation. Experimental results demonstrate that the proposed method achieves an accuracy of 97.82% on the apple disease classification task, while obtaining MPA and MIoU scores of 94.85% and 91.75%, respectively, on the segmentation task.
Early detection and characterisation play a crucial role in effectively managing and treating chronic illness. Given the increasing prevalence of chronic kidney disease (CKD), the healthcare system faces a significant...
详细信息
The early discovery of cervical cancer is crucial for efficient treatment and increased survival rates, making it a severe public health concern [1]. This study uses a consistent dataset to compare various machine-lea...
详细信息
ISBN:
(纸本)9798350351194;9798350351187
The early discovery of cervical cancer is crucial for efficient treatment and increased survival rates, making it a severe public health concern [1]. This study uses a consistent dataset to compare various machine-learning methods for cervical cancer prediction. We utilized a variety of machinelearning techniques, including Random Forest, Naive Bayes, Support Vector machine (SVM) with a linear kernel, K-Nearest Neighbors (KNN), Logistic Regression, and Extreme Gradient Boosting (XGBoost), to identify and forecast the risk of cervical cancer. Based on the accuracy, precision, recall, F1-score, and confusion matrices, the effectiveness of these algorithms was assessed [2]. The most appropriate model for this application is XGBoost, which fared better than other models in recall and F1-score, even if more conventional methods, such as Random Forest and KNN, showed excellent overall accuracy. The study results imply that XGBoost has excellent potential for creating an efficient cervical cancer screening tool due to its balance of sensitivity and precision. The model is then integrated into a web-based application and an interactive chatbot designed to facilitate early detection and assessment of cervical cancer risks.
This is increasingly important as global obesity rates continue to rise: the causes need to be understood, and the ability to predict trends determined. In this study, we use machinelearning techniques to predict obe...
详细信息
Various crowd models developed in the field of crowd simulation have been used to learn action policies of reinforcement learning for the navigation of autonomous mobile robots. However, there is no single crowd model...
详细信息
ISBN:
(纸本)9798350358513;9798350358520
Various crowd models developed in the field of crowd simulation have been used to learn action policies of reinforcement learning for the navigation of autonomous mobile robots. However, there is no single crowd model that can be applied to all environments because crowd movement is highly dependent on the environment. In this paper, we propose a method for selecting an appropriate crowd model for each environment by classifying the spatio-temporal data to be simulated into multiple categories. For that, we extract features related to crowd movement and geographic shape for use in the robot simulation. Specifically, we generate image data to visualize pedestrian movement trends, perform feature extraction using an autoencoder, and classify the results into several categories. Through evaluation experiments using pedestrian movement trajectory datasets, this paper shows that visually similar feature images are classified into the same categories.
Our research proposes a comprehensive approach to identify duplicate frames in digital videos. It integrates machinelearning and signal processing techniques for effective identification. The process begins with pre-...
详细信息
Cutting-edge technology is essential to improve resource use, output, and farming methods to meet climate change and other needs. This research uses IoT and machinelearning to remotely track and operate agricultural ...
详细信息
Cutting-edge technology is essential to improve resource use, output, and farming methods to meet climate change and other needs. This research uses IoT and machinelearning to remotely track and operate agricultural *** Internet of Things (IoT) lets several sensors strategically placed around the farm collect data simultaneously. These sensors assess temperature, humidity, and soil wetness, which determine crop health. The data is wirelessly delivered from a central location to a designated area for processing and *** learning is used to interpret data. These algorithms can improve irrigation schedules, predict food yields, diagnose diseases, and offer insect control alternatives. machinelearning (ML) models improve by learning from data and adapting to environmental *** system allows remote farming monitoring using actuators and automated tools. An easy-to-use interface on desktops or mobile devices allows agricultural experts utilise machinelearning models to regulate insecticides, modify irrigation water, and start harvesting *** proposed technology improves farming productivity, durability, resource efficiency, and cost. It encourages farmers to make sensible decisions and manage their resources to better address environmental *** conclude, machinelearning and the Internet of Things could considerably improve farming system remote monitoring and control. The development of new technologies in this field is crucial to its long-term success and ability to provide for future generations.
Clinical patients should have corresponding clinical indicators to characterize the disease. Multimodal data and physiological indicators provide a basis for patient diagnosis and assessment, but small sample data pos...
详细信息
ISBN:
(纸本)9798400716379
Clinical patients should have corresponding clinical indicators to characterize the disease. Multimodal data and physiological indicators provide a basis for patient diagnosis and assessment, but small sample data pose statistical difficulties. In order to better support the clinical conclusions, from a data-driven perspective, using machinelearning algorithms, we explored the support of physiological indicator data for multimodal data in the case of insufficient samples, and according to the results of the model, it is shown that the data-driven results can better support the final conclusions, and therefore, integrating the multimodal data and the clinical indicators can better provide the diagnosis and assessment conclusions for the clinical patients.
This paper delves into the significance of predictive maintenance, a technique leveraging intelligent sensors for efficient data collection and evaluation to prevent problems and accidents. A data-focused predictive m...
详细信息
Early detection of stuck pipes during drilling operations is crucial and challenging. Some of the existing studies on the stuck pipe detection have adopted supervised machinelearning approaches that employ datasets f...
详细信息
ISBN:
(纸本)9780791887868
Early detection of stuck pipes during drilling operations is crucial and challenging. Some of the existing studies on the stuck pipe detection have adopted supervised machinelearning approaches that employ datasets for "stuck" and "normal". However, for early detection before the occurrence of stuck pipe, the application of ordinal binary classification in supervised machinelearning has presented several elemental concerns, such as limited stuck pipe data, and lack of an exact "stuck sign" before occurrence. Our previous studies proposed unsupervised machinelearning approaches using data on the normal activities with long short-term memories and autoencoder, the mixture probability model, and graph attention model, and presented the possibility of predicting a stuck pipe. On the other hand, there observed false positives responding the operation of drilling equipment or changes in drilling data. We introduce a datascience approach, including machinelearning, incorporating physical knowledge to overcome frequent false positives. This study initially introduces the stuck pipe predictions using unsupervised machinelearning approaches, and a typical false positive case. Subsequently, we present machinelearning integrating physical knowledge and demonstrate early stuck pipe detection using field data containing stuck pipe events. The results show significant suppression of false positives. Our machinelearning approach is expected to contribute to considerable reduction of nonproductive time in drilling operations, potentially preventing well abandonment.
暂无评论