A hybrid bio-inspired augmented with hyper-parameter deep learning model for brain tumor classification
Abstract Labeled data are required for the development of powerful deep learning (DL) models for medical imaging diagnoses. Because learning from such large datasets is difficult, medical imaging data analysis is becoming increasingly popular employing bio-inspired enabled deep learning models. As a...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2025-07-01
|
| Series: | Journal of Electrical Systems and Information Technology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s43067-025-00207-y |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Labeled data are required for the development of powerful deep learning (DL) models for medical imaging diagnoses. Because learning from such large datasets is difficult, medical imaging data analysis is becoming increasingly popular employing bio-inspired enabled deep learning models. As a result, this research presents a DL-based model according to brain tumor (BT) classification bio-inspired principles. The suggested approach uses a mix of bio-inspired optimization approaches and convolutional neural networks (CNNs) to classify brain tumors. The CNN model is adjusted for different convolutional layers and fully connected layers to identify patterns and features in brain tumor pictures using an enhanced salp swarm algorithm (SSA) with kernel extreme learning machine (KELM). To further improve the performance of the model, SSA was increased to select relevant features. This study examines how SSA may be employed for feature selection thereby increasing the efficiency of the DL models using hybrid methods. Brain tumor datasets were employed in assessing the efficacy of the suggested model. The study utilizes the CE-MRI Figshare dataset containing 3064 T1-weighted contrast MRI slices from 233 patients diagnosed with meningioma, glioma, and pituitary brain tumors for training and testing the proposed model. An accuracy of 99.9%, 99.5% sensitivity, 99.9% specificity, and 99.5% F1-score was obtained from the experimental study. The suggested model classified the brain tumors efficiently as proved by the execution time of 0.089 s. The suggested model outperformed the baseline CNN models according to the datasets. The proposed model can be employed as a diagnostic tool for other medical imaging applications as evident in the study’s findings. There is a significant improvement in the accuracy and efficiency of brain tumors classification utilizing the proposed model. |
|---|---|
| ISSN: | 2314-7172 |