You are on page 1of 17

BRAIN TUMOR DETECTION

USING
Presented By: CNN & DEEP LEARNING
R Poojitha-206W1A4356
M NandiniDevi-206W1A4343 Under the Esteemed Guidance of:
N Kavya-206W1A4345 B.Susmitha M.Tech
CH Gayatri-206W1A4314 Assistant Professor
Agenda
• Introduction
• Abstract
• Existing System
• Disadvantages of Existing System
• Proposed System
• Flow Diagram
• Understanding Brain Tumors
• Software Requirement Specification
• Hardware Requirement Specification
Abstract
• A diverse and well-annotated dataset of brain MRI scans,
encompassing various tumor types and healthy brain tissues,
forms the foundation of our research.
• We propose a CNN architecture tailored for brain tumor
detection, emphasizing the extraction of hierarchical features and
spatial relationships within the medical images.
• The model undergoes rigorous training on the dataset, facilitating
its ability to recognize complex patterns associated with different
tumor categories.
INTRODUCTION
Brain tumors are a critical health concern,
necessitating accurate and timely diagnosis for
effective treatment. With advancements in medical
imaging and artificial intelligence, there has been a
notable shift towards leveraging Convolutional
Neural Networks (CNNs) and Deep Learning (DL)
techniques for the automated detection and
classification of brain tumors. This integration holds
immense potential to enhance diagnostic accuracy,
reduce human error, and expedite treatment
decisions.
Existing System
 DL enables a pre-trained Convolutional Neural Network
(CNN) model for medical images, specifically for classifying
brain cancers.
 As MRI images are prone to more noise and interference,
doctors felt it difficult to detect the tumor at an early stage.
 They not only felt it difficult to detect the tumor at an early
stage, they also took many days to detect it manually.
Disadvantages of Existing System
 Limited Contrast Resolution: MRI and CT scans may struggle to
distinguish tumors from healthy tissue, but image
enhancement techniques like contrast stretching improve
visibility.
 Presence of Noise and Artifacts: Noise and artifacts in MRI and
CT images can obscure details, but using denoising algorithms
enhances image clarity for better tumor detection.
 Variability in Image Quality: Inconsistent image quality across
scans can complicate tumor detection, yet standardizing
protocols and applying quality enhancement algorithms
ensures more reliable results.
Proposed System
• The proposed work is to overcome the existing system.
• This System detects the tumor from the MRI images through
image processing methods like Noise Reduction, and contrast
Enhancement that include some techniques.
• Those techniques are the modules of the project.
Block Diagram

KNN
Image Enhancement
The Histogram equalization is used to enhance the quality and
contrast of image.
Skull removal
Cranium in the MRI has approximately same intensity as the tumor
and it can interfere in the segmentation and thresholding step hence
removing the skull in the enhanced image is essential.
KNN means segmentation

KNN, as a soft classification technique, allows for flexible assignment of


data points to clusters, which is advantageous in image segmentation tasks
where ambiguity and information retention are important. It is particularly
popular in image segmentation due to its ability to handle complex image
data and provide detailed segmentations with smooth transitions between
regions.
KNN segmentation
Morphological Operations
Morphological operations like erosion and dilation are used to
remove these white components.
Tumor Extraction
Software Requirement
Specification

Matplotlib-Matplotlib helps
with data analyzing,and is a
numerical plotting
library
Hardware Requirement
Specification

CPU: 16 cores or more with high clock


speeds (3.5 GHz or higher) or consider
server-grade CPUs with even higher
performance.
RAM: 4GB or more is recommended,
depending on model complexity and
dataset size.
Thank You

You might also like