Medical Image Analysis (MIA) is integral to healthcare, demanding advanced computational techniques for precise diagnostics and treatment planning. The demand for accurate and interpretable models is imperative in the...
详细信息
Medical Image Analysis (MIA) is integral to healthcare, demanding advanced computational techniques for precise diagnostics and treatment planning. The demand for accurate and interpretable models is imperative in the ever-evolving healthcare landscape. This paper explores the potential of Self-Supervised Learning (SSL), transfer learning and domain adaptation methods in MIA. The study comprehensively reviews SSL-based computational techniques in the context of medical imaging, highlighting their merits and limitations. In an empirical investigation, this study examines the lack of interpretable and explainable component selection in existing SSL approaches for MIA. Unlike prior studies that randomly select SSL components based on their performance on natural images, this paper focuses on identifying components based on the quality of learned representations through various clustering evaluation metrics. Various SSL techniques and backbone combinations were rigorously assessed on diverse medical image datasets. The results of this experiment provided insights into the performance and behavior of SSL methods, paving the way for an explainable and interpretable component selection mechanism for artificial intelligence models in medical imaging. The empirical study reveals the superior performance of BYOL (Bootstrap Your Own Latent) with resnet as the backbone, as indicated by various clustering evaluation metrics such as Silhouette Coefficient (0.6), Davies-Bouldin Index (0.67), and Calinski-Harabasz Index (36.9). The study also emphasizes the benefits of transferring weights from a model trained on a similar dataset instead of a dataset from a different domain. Results indicate that the proposed mechanism expedited convergence, achieving 98.66% training accuracy and 92.48% testing accuracy in 23 epochs, requiring almost half the number of epochs for similar results with ImageNet weights. This research contributes to advancing the understanding of SSL in MIA, providin
In serverless computing, the service provider takes full responsibility for function management. However, serverless computing has many challenges regarding data security and function scheduling. To address these chal...
详细信息
This paper centers on leveraging Convolution-Augmented Transformer Models originally designed for Automatic Speech Recognition (ASR) in the realm of Sign Language—specifically, American Sign Language Fingerspelling R...
详细信息
This paper presents MACPGANA, a novel multimodal agriculture commodity price prediction model leveraging Generative Adversarial Networks (GAN) and Autoencoders (AEs). The model integrates multimodal data to extract an...
详细信息
Age-related macular degeneration is a chronic disease affecting a central area of the retina. Accurate disease identification aids in slowing down the progression of age-related macular degeneration and preserving vis...
详细信息
Age-related macular degeneration is a chronic disease affecting a central area of the retina. Accurate disease identification aids in slowing down the progression of age-related macular degeneration and preserving vision. Various traditional techniques have been developed for effective age-related macular degeneration detection. However, traditional approaches failed to detect and classify the disease accurately and it consumes more time. However, traditional approaches failed to detect and classify age-related macular degeneration accurately. This research paper proposed an efficient model named as Multi-Modal Vision transformer model for the early and accurate prediction of age-related macular degeneration. This study aims to combine information from the Color Fundus Photography and Optical Coherence Tomography streams for performing efficient age-related macular degeneration diagnosis. The input images are needed to be preprocessed to enhance the image quality and make it suitable for further processing. The proposed framework integrated a Cascaded group attention transformer block which extracts the significant features from these modalities effectively. This block has the ability to solve computational complexity issues and attention head redundancy problems. Further, the multi-modal fusion method based on self-attention is introduced for fusing the features from Color Fundus Photography and Optical Coherence Tomography images. This fusion model is trained by applying both standard backpropagation and random gradient descent algorithms. For multi-class classification tasks, the fused features are classified into different classes based on the decision score. To visualize the single-modal and multi-modal output images in a heat map we applied a Class Activation Mapping model. Furthermore, the proposed technique is conducted on the Python platform and the performance is evaluated on different datasets with significant evaluation measures. This technique achieves
This paper proposes a Poor and Rich Squirrel Algorithm (PRSA)-based Deep Maxout network to find fraud data transactions in the credit card system. Initially, input transaction data is passed to the data transformation...
详细信息
Voice is the king of communication in wireless cellular network (WCN). Again, WCNs provide two types of calls, i.e., new call (NC) and handoff call (HC). Generally, HCs have higher priority than NCs because call dropp...
详细信息
Cancer remains a leading cause of mortality worldwide, with early detection and accurate diagnosis critical to improving patient outcomes. While computer-aided diagnosis systems powered by deep learning have shown con...
详细信息
The core task in natural language processing (NLP) is text summarization, which condenses important information from large volumes of text into brief summaries. This study reviews text summarization strategies using N...
详细信息
作者:
Mahapatra, AbhijeetPradhan, RosyMajhi, Santosh K.Mishra, Kaushik
Department of Computer Science & Engineering Odisha Burla768018 India Sikkim Manipal University
Sikkim Manipal Institute of Technology Department of Artificial Intelligence and Data Science Sikkim India
Department of Electrical Engineering Odisha Burla768018 India
Department of Computer Science and Information Technology Chhattisgarh Bilaspur495009 India Manipal Academy of Higher Education
Manipal Institute of Technology Bengaluru Department of Computer Science and Engineering Manipal India
The rapid proliferation of IoT devices like smartphones, smartwatches, etc. has significantly elevated the quantity of data requiring execution. It poses challenges for centralized Cloud computing servers, such as lat...
详细信息
暂无评论