IoT based ECG Signal Feature Extraction and Analysis for Heart Disease Risk Assessment

Satheeskumaran, S.; Sasikala, K.; Kumar, Neeraj; Senthilkumar, A.; Babu, N. Sharath

doi:10.1109/icscss57650.2023.10169414

Cited by 7 publications

(1 citation statement)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The outcomes, which are shown in Table 4, demonstrate how well the ResNet50-CBAM performed in comparison to other methods. Importantly, as Table 4 shows, the training and evaluation techniques from these previous publications were used to evaluate the ResNet50-CBAM model [39]. A elimination study was also performed on the model using the additional parameter values and the 80% and 20% train test split estimation approach.…”

Section: Resultsmentioning

confidence: 99%

Enhanced Brain Tumor Analysis: Integrating ResNet50 with Convolutional Block Attention Modules for Advanced Insights

Vinston Raja R.

2024

jes

View full text Add to dashboard Cite

The process of diagnosing brain tumors is a lengthy task that heavily relies on the experience of radiologists. However, deep learning techniques have become increasingly popular in automating the diagnosis of brain tumors, offering improved precision and effectiveness. One such technique, Convolutional Block Attention Modules (CBAM), uses attention-based models that dynamically enhance and refine diagnostic characteristics. However, the specific impact of using different attention methods, such as channel, spatial, or combined attention, within CBAM for brain tumor classification is yet to be fully explored.To address this gap, our research used ResNet50 coupled with CBAM to classify brain tumors. This novel approach demonstrated superior performance compared to existing methods, including Convolutional Neural Network. ResNet50-CBAM showed remarkable area under the curve (AUC), recall, precision, and accuracy of 99.53%, 99.11%, 98.75%, and 99.35%, respectively, using the same dataset.The fusion of ResNet-CBAM not only captures spatial context but also enhances feature representation, making it a promising integration into brain classification software platforms. This development could benefit doctors by improving brain tumor categorization and facilitating better clinical decision-making.

show abstract

Section: Resultsmentioning

confidence: 99%