Semantic-Aware Hybrid Deep Learning Model for Brain Tumor Detection and Classification Using Adaptive Feature Extraction and Mask-RCNN

Semantic-Aware Hybrid Deep Learning Model for Brain Tumor Detection and Classification Using Adaptive Feature Extraction and Mask-RCNN

Anil Kumar Mandle (Department of Information Technology, National Institute of Technology Raipur, India), Govind P. Gupta (Department of Information Technology, National Institute of Technology Raipur, India), Satya Prakash Sahu (Department of Information Technology, National Institute of Technology Raipur, India), Shavi Bansal (nsights2Techinfo, India & Center for Interdisciplinary Research, University of Petroleum and Energy Studies (UPES), Dehradun, India), and Wadee Alhalabi Alhalabi (Department of Computer Science, Immersive Virtual Reality Research Group, King Abdulaziz University, Jeddah, Saudi Arabia)
Copyright: © 2025 |Pages: 23
DOI: 10.4018/IJSWIS.365910
Article PDF Download
Open access articles are freely available for download

Abstract

A brain tumor is one of the most prevalent causes of cancer death. The best strategy is the timely treatment of brain tumors in their early detection. Magnetic Resonance Imaging (MRI) is a standard non-invasive method to detect brain tumors. For early detection and better patient survival through MRI scans, the diagnosis needs a high level of knowledge in the radiological and neurological domains to identify the cancers. Researchers have suggested various brain cancer detection techniques. However, most existing automatic cancer detection approaches suffer from poor accuracy and low detection rates. This paper proposes a hybrid deep learning (DL) using deep feature extraction and adaptive Mask Region-based Convolutional Neural Networks (Mask-RCNNs) model for brain tumor detection and classification method to overcome these issues. The experimental findings on the benchmark dataset demonstrate that the planned model is highly effective, with 99.64% accuracy, 95.93% precision, 95.39% recall, and 95.67% F1-score.
Article Preview
Top

Introduction

A brain tumor, whether benign or malignant, develops on the brain or skull walls. The growth of skull tumors can stress the brain, resulting in a negative effect on overall health. Timely identification of brain tumors is, therefore, a significant area of investigation within medicinal imaging to identify the most effective treatment options to save a patient’s life.

Over the last few decades, medical scientists have discovered more than 120 various forms of brain cancer. These are divided into primary brain cancers, which develop in the brain cells, and secondary brain cancers, which form in the surrounding organs and spread to the brain through blood circulation (Anaraki et al., 2019; Behin et al., 2003). Primary cancers account for approximately 70% of all brain malignancies; secondary cancers account for 30% (Sultan et al., 2019). According to the National Brain Tumor Foundation, about 29,000 instances of primary brain cancers are identified every year in the United States, with 13,000 deaths each year (Akil et al., 2020; Nhi et al., 2022). More than 42,000 people in the United Kingdom die from primary brain cancers each year.

The significant categories of brain cancers include glioma (G), meningioma (M), and pituitary (P). Glioma has the uppermost death and flexibility rates of the different cancers (Maharjan et al., 2020). Low-grade gliomas and high-grade gliomas are two types of gliomas that arise from glial cells in the brain. High-grade gliomas are dangerous, with the patient generally receiving a two-year persistence rate (Qian et al., 2022; Smoll et al., 2013). Meningioma usually forms on the protective tissue layer covering the brain and spinal cord. This type of cancer is often less dangerous and slow in growth (Chu et al., 2022; Louis et al., 2016). Cancers near the pituitary gland, which synthesizes important hormones, are usually benign. However, pituitary cancers affect the functioning of this gland (Masood et al., 2021), threatening the patient’s life if not diagnosed in a timely manner.

Various medical imaging modalities can be employed for cancer diagnosis in clinical settings, with the choice depending on the situation and objectives (Chui et al., 2023; Komninos et al., 2004). Non-invasive imaging procedures like computed tomography (CT), MRI, and positron emission tomography (PET) are preferred for early-stage brain tumor detection over invasive methods like biopsies (Ker et al., 2017).

An MRI image is the most effective intra-operative collective imaging process because it does not use hazardous ionizing radiation like x-rays. It provides risk-free, high-quality images of soft tissue, as well as the ability to collect modalities with several parameters, including T1, T1c, T2, and FLAIR. Each modality generates a distinct muscle contrast (Bauer et al., 2013; Chui et al., 2022). The fundamental objective of early therapy for a neurosurgeon is to identify cancer sites correctly. Otherwise, excessive or insufficient excision may cause discomfort or irreversible loss. Manual brain cancer segmentation requires extensive domain knowledge and significant time investment. In addition, it is subject to inter- and intra-variability observations (Masud et al., 2021; Olabarriaga & Smeulders, 2001). Computer-aided diagnosis methods are increasingly utilized as assistive tools to aid in the recognition and classification of brain cancers in MRI scans.

Modern studies illustrate vital advancements in semi-automatic or automatic cancer segmentation techniques. However, research on accurate cancer segmentation remains a challenge (Asa, 1998). First, cancer size, location, and appearance vary from patient to patient (Işın et al., 2016). Second, malignant margins may be abnormal because healthy tissue often fills malignant areas (Goetz et al., 2015). Third, low signal-to-noise ratio or image bias may be caused by MRI acquisition techniques or imaging devices (Mamta & Gupta, 2021; Yao, 2006).

Complete Article List

Search this Journal:
Reset
Volume 21: 1 Issue (2025)
Volume 20: 1 Issue (2024)
Volume 19: 1 Issue (2023)
Volume 18: 4 Issues (2022): 2 Released, 2 Forthcoming
Volume 17: 4 Issues (2021)
Volume 16: 4 Issues (2020)
Volume 15: 4 Issues (2019)
Volume 14: 4 Issues (2018)
Volume 13: 4 Issues (2017)
Volume 12: 4 Issues (2016)
Volume 11: 4 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing