Deep Learning Approach to Classify Brain Tumor with Comparative Analysis of CT and MRI Scans
Journal Title: International Journal of Innovative Research in Computer Science and Technology - Year 2022, Vol 10, Issue 6
Abstract
Brain tumor is an intracranial growth or collection of aberrant cells. While brain tumors can afflict anyone at any age, they most typically affect youngsters and the elderly. The aberrant tissue cells in brain tumors are notoriously challenging to classify because of the variety of these malignancies which negatively affect human health and jeopardize life. Therefore, early detection of aberrant features is essential for tumor treatment. The motivation to do this research is to enhance the competency in terms of accuracy, speedy detection and less validation loss by employing CNN (7x7 matrix). Brain CT imaging is typically the first radiologic test performed when a tumor is suspected. However, MRI offers very good soft tissue characterization capabilities along with high quality images. This manuscript includes the comparison of CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) images for the diagnosis of brain tumor. The proposed study work utilizes a CNN-based model and min-max normalization to divide 7023 and 3249 T1-weighted contrast enhanced brain MRI and CT SCAN pictures into four groups (glioma, meningioma, pituitary, and no tumor). Photos of tumors from medical records are used in the suggested strategy based on computer-aided diagnostic research. It introduces the segmentation and classification of tumor images as well as the diagnosis approaches based on CNN to help clinicians recognize cancers. This new network, which incorporates drop-out and dense layers, is an adaption of CNN wherein data augmentation with min-max normalization and six convolutional layers are employed to enhance the contrast of tumor cells using Kaggle dataset. The experimental results show that the proposed model was validated to obtain 97.78% accuracy and 0.087 validation loss during testing and training using medical imaging techniques with precision. The model's overall efficiency was raised by employing 10 epochs.
Authors and Affiliations
Zarka Ashraf, Ravinder Pal Singh, and Dr. Monika Mehra
Keywords
Related Articles
- EP ID EP746017
- DOI 10.55524/ijircst.2022.10.6.17
- Views 35
- Downloads 0
An Experimental Study on Soil Stabilization by Using Cement and Incinerated Bottom Ash
This main objective of this study is to improve the strength of peat soil by adding industrial waste as a filler and cement as a binder. Peat soil are known to have a weak strength and a very high settlement problem. Ho...
Deep Learning-based Classification of Materials into Biodegradable and Non-Biodegradable
The collection of real-time data from people, their cars, public transit, buildings, and other urban infrastructures like the energy grid and waste management systems are at the heart of the smart city concept. The insig...
An Induction Generator based AC/DC Hybrid Electric Power Generation System for Additional Electrically Powered Airliner
For an additional aircraft power system, both AC and DC power having high voltage levels are required. with various flight loads. This paper introduces an induction generator based on AC / DC hybrid MEA power generation...
A Centralised Group Aware Duty Cycle Control Protocol for Wireless Sensor Networks
Wireless sensor networks are collection of autonomous devices comprising of sensors to sense the effects of temperature, pressure or pollutions in our environment at different locations in real time. These devices may be...
Evolving Constraints in Military Applications using Wireless Sensor Networks
WSNs consist of a large number of small sensor nodes. These nodes are very cheap in terms of cost. In military operations, there is always a threat of being attacked by enemies. So, the use of these cheap sensor nodes wi...