Year
Publisher
Journal
1
Institution
Institution Country
Publication Type
Field Of Study
Topics
Open Access
Language
Filter 1
Year
Publisher
Journal
1
Institution
Institution Country
Publication Type
Field Of Study
Topics
Open Access
Language
Filter 1
Export
Sort by: Relevance
M-C&M-BL: a novel classification model for brain tumor classification: multi-CNN and multi-BiLSTM

As the central organ for human cognition and behavior, the human brain is critical to daily functioning. Brain tumors disrupt normal activities and require accurate diagnosis and intervention. In this study, an approach is presented that allows the detection of tumors in the brain. The main motivation of the study is to determine whether there is a tumor in the brain with high performance. For this purpose, the proposed method was tested using the open-source Br35H brain magnetic resonance imaging (MRI) dataset. The proposed model, M-C&M-BL, integrates a Convolutional Neural Network (CNN) for image feature extraction with a Bidirectional Long Short-Term Memory (BiLSTM) Network for sequential data processing. Metrics such as Accuracy (Acc), F1 Score (F1), Precision (Pre), Recall (Rec), Specificity (Spe), and Matthews Correlation Coefficient (MCC) were used for performance evaluation. The proposed model achieved 99.33% Acc and 99.35% F1, outperforming CNN-based models such as BMRI-Net (98.69% Acc, 98.33% F1) and AlexNet (98.79% Acc, 98.82% F1). It also demonstrated competitive performance against MobileNetv2, which achieved a slightly higher Acc of 99.67%. This approach has significant potential for integration into clinical decision support systems, web and mobile diagnostic platforms, and hospital picture archiving and communication systems (PACS). These tools can aid in early diagnosis, improve diagnostic accuracy, and reduce evaluation time. However, challenges such as ensuring privacy, achieving generalizability across diverse datasets, and addressing infrastructure constraints must be addressed for seamless deployment. This study highlights the feasibility and potential of combining deep learning architectures to advance AI-driven tools in healthcare, ultimately improving clinical workflows and patient outcomes.

Read full abstract
Open Access Just Published