Brain Tumor Classification Using FCM, DenseNet, and SVM with Transfer Learning Techniques
Brain tumor classification is a critical task in medical imaging, aimed at distinguishing between different tumor types for timely diagnosis and treatment planning. This study presents a hybrid approach that combines Fuzzy C-Means (FCM) clustering, DenseNet architecture, and Support Vector Machine (...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
EDP Sciences
2025-01-01
|
| Series: | EPJ Web of Conferences |
| Online Access: | https://www.epj-conferences.org/articles/epjconf/pdf/2025/13/epjconf_icetsf2025_01035.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Brain tumor classification is a critical task in medical imaging, aimed at distinguishing between different tumor types for timely diagnosis and treatment planning. This study presents a hybrid approach that combines Fuzzy C-Means (FCM) clustering, DenseNet architecture, and Support Vector Machine (SVM) for accurate brain tumor classification. Initially, FCM is applied to preprocess MRI images, enhancing tumor regions for better feature extraction. The preprocessed images are then fed into a DenseNet model, leveraging transfer learning to extract deep features from the tumor regions. DenseNet, known for its dense connections between layers, improves feature propagation and reduces vanishing gradient issues, allowing for efficient training on medical imaging datasets. Finally, the extracted features are classified using an SVM classifier, which is effective in handling high-dimensional data and separating classes with a maximal margin. The proposed method aims to enhance classification accuracy by integrating the strengths of FCM for image preprocessing, DenseNet for feature extraction, and SVM for classification. Experimental results demonstrate the effectiveness of this approach in achieving high accuracy and robustness compared to traditional methods. This framework holds potential for improving the early diagnosis of brain tumors, aiding in better patient outcomes through precise classification. |
|---|---|
| ISSN: | 2100-014X |