Photography is the most important, powerful, and reliable means of expression. Today, digital images not only provide disinformation but also act as agents for secret communication. Users and editing professionals wor...
详细信息
Empirical robustness evaluation (RE) of deep learning models against adversarial perturbations involves solving non-trivial constrained optimization problems. Recent work has shown that these RE problems can be reliab...
详细信息
The building extraction of the footprint from the satellite imagination has been a research problem that needs to be solved efficiently. The hybrid semantic segmentation framework is used to increase the footprint ext...
详细信息
With the static nature of recommender systems, it is crucial to take into account the intricate details of customer behaviour to better improve operating efficiencies. This paper is about original methods to augment r...
详细信息
Air pollution is a significant environmental hazard in modern society because of its serious impact on human health and the environment. In point of fact, there has been a substantial rise in the levels of pollution i...
详细信息
In a rapidly urbanizing world, significant environmental changes such as deforestation, climate change, and disasters are occurring swiftly, leading to critical land use and land cover (LULC) shifts. Although survey d...
详细信息
One of the prominent causes of death for women universally is breast cancer. The number of early deaths is reduced with early detection. Using an ultrasound scan, the data examine the ultrasound images of breast cance...
详细信息
Automated dermatological diagnosis with artificial intelligence has made significant progress in recent years, especially with deep learning algorithms. Despite the advancement, classifying rare and unseen skin lesion...
详细信息
ISBN:
(纸本)9781510683327
Automated dermatological diagnosis with artificial intelligence has made significant progress in recent years, especially with deep learning algorithms. Despite the advancement, classifying rare and unseen skin lesion types remain a challenge mainly due to the limited availability of annotated data. In this paper, we propose a new approach that combines a transfer learning algorithm with zero-shot learning using the HAM10000 dataset. We use a pre-trained VGG-16 model with ImageNet weights to extract features from dermatoscopic images, finetune it on a subset of common lesion types in the dataset, and then use zero-shot learning principles to infer diagnosis for rare or unseen lesion types. This is done by using semantic embeddings from the penultimate layer of the network and leveraging similarity metrics like cosine similarity to compare embeddings of unseen lesions with prototypes of known lesion types. The experimental methodology involves training the model on a labeled subset of HAM10000, validating on a held-out test set, and testing on unseen lesion types. We use multiple evaluation metrics namely AUC, accuracy, precision, recall, and F1-score to evaluate the model's performance across different lesion categories. Our model achieves an overall accuracy of 93% and AUC of 0.87 on the test set. We also explore the robustness of the approach against class imbalance and variability in lesion morphology and quality. Our experimental result shows that the combined approach performs better than the traditional transfer learning methods, especially for rare and less frequent lesion types. The results demonstrate the potential of combining transfer learning with zero-shot learning to improve the precision, generalization, and clinical utility of automated dermatological diagnosis systems. This proposed approach could potentially overcome the limitations of current dermatological imaging techniques and lay the foundation for the development of more effective and dependa
Data mining process involves a number of steps fromdata collection to visualization to identify useful data from massive data *** same time,the recent advances of machine learning(ML)and deep learning(DL)models can be...
详细信息
Data mining process involves a number of steps fromdata collection to visualization to identify useful data from massive data *** same time,the recent advances of machine learning(ML)and deep learning(DL)models can be utilized for effectual rainfall *** this motivation,this article develops a novel comprehensive oppositionalmoth flame optimization with deep learning for rainfall prediction(COMFO-DLRP)*** proposed CMFO-DLRP model mainly intends to predict the rainfall and thereby determine the environmental ***,data pre-processing and correlation matrix(CM)based feature selection processes are carried *** addition,deep belief network(DBN)model is applied for the effective prediction of rainfall ***,COMFO algorithm was derived by integrating the concepts of comprehensive oppositional based learning(COBL)with traditional MFO ***,the COMFO algorithm is employed for the optimal hyperparameter selection of the DBN *** demonstrating the improved outcomes of the COMFO-DLRP approach,a sequence of simulations were carried out and the outcomes are assessed under distinct *** simulation outcome highlighted the enhanced outcomes of the COMFO-DLRP method on the other techniques.
At present, there exist some problems in granular clustering methods, such as lack of nonlinear membership description and global optimization of granular data boundaries. To address these issues, in this study, revol...
详细信息
暂无评论