Optimized Fuzzy Logic Adaptive System with Holographic Convolutional Heterogeneous Graph Neural Network-based Feature Extraction and Classification for Brain Tumor Detection

Cells growing abnormally in the brain or surrounding tissues is called a brain tumor. It can cause symptoms such as headaches, seizures, and physical or cognitive deficits. It can also be benign (non-cancerous) or malignant (cancerous). To control its effects and enhance results, early identificatio...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (Online) s. 1324 - 1330
Hlavní autoři: Regan, M., Srinivasan, P. Santhosh
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 03.10.2024
Témata:
Accuracy
Adaptive systems
Brain Tumor Detection
Classification algorithms
Convolutional neural networks
Feature extraction
Filter based Fuzzy Logic Adaptive System
Fuzzy logic
Graph neural networks
Heterogeneous Graph Neural Network Boosted Sooty Tern Optimization Algorithm
Holographic Convolutional Neural Network
Optimization
Real-time systems
Robustness
ISSN:2768-0673
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Cells growing abnormally in the brain or surrounding tissues is called a brain tumor. It can cause symptoms such as headaches, seizures, and physical or cognitive deficits. It can also be benign (non-cancerous) or malignant (cancerous). To control its effects and enhance results, early identification and treatment are essential. It is necessary to detect BC, because of that many techniques are implemented. But the existing methods have lot of disadvantages such as, overfitting high computational complexity and resource demands. To overcome the aforementioned problem, Holographic Convolutional Neural Network with Heterogeneous Graph Neural Networks using Boosted Sooty Tern Optimization Algorithm (HCNN-HGNN-BSTOA) is proposed for accurately detecting BC. . In this input image is taken from Figshare dataset and Brats2020 dataset .The pre-processing and feature extraction are done using Filter based Fuzzy Logic Adaptive System Based Feature Extraction (CFFLAS) is proposed, for the purpose of removing noises and improve the efficiency of feature extraction. Following that future selection are done using Fennec Fox Optimization Algorithm (FFOA). After that classification are done using Holographic Convolutional Neural Network (HCNN)and optimization are done using Boosted Sooty Tern Optimization Algorithm (BSTOA) for detecting normal and abnormal conditions of the BT. The efficiency of the proposed HCNN-HGNN-BSTOA is analyzed using an dataset and attains 99.78% accuracy, 97.34% recall and attains better results compared with the existing methods. Experimental results demonstrate the effectiveness of the proposed approach in accurately detecting tumors in medical images, paving the way for improved diagnostic tools in healthcare.
ISSN:2768-0673
DOI:10.1109/I-SMAC61858.2024.10714877