Spring 2024 - CS619 - 9340

NeuroScan: Advanced Brain Tumor Detection System
Project Domain / Category

Image Processing/Deep Learning
Abstract / Introduction
NeuroScan is an innovative project aimed at enhancing the efficiency and accuracy of brain tumor detection
using advanced machine learning and medical imaging techniques. With the increasing incidence of brain
tumors worldwide, early and accurate diagnosis is crucial for timely intervention and treatment planning.
NeuroScan seeks to address this challenge by developing a robust and reliable system for automated brain
tumor detection.
The admin (student) will develop a system (NeuroScan) utilizing deep learning algorithms trained on large
datasets of medical images, including MRI (Magnetic Resonance Imaging) scans and CT (Computed
Tomography) scans. The designed system (NeuroScan) will recognize patterns and anomalies indicative of
brain tumors with high precision and sensitivity.
Functional Requirements:
Admin (Student) will perform all the following tasks.
1. Image Input: NeuroScan should support the input of medical images from MRI and CT scans. The
system should be capable of handling various resolutions and image qualities to accommodate
different imaging devices.
2. Preprocessing: NeuroScan must preprocess input images to remove noise, artifacts, and irrelevant
structures that could interfere with tumor detection. Preprocessing techniques should include image
normalization, noise reduction, Skull Stripping, and artifact removal.
3. Data Augmentation: NeuroScan shall facilitate the augmentation of training data to enhance the
diversity and robustness of the dataset utilized for brain tumor detection. The system shall offer a suite
of data augmentation techniques, including rotation, translation, scaling, flipping, elastic deformation,
intensity adjustment, noise injection, shearing, and random cropping.
Apply the Data Augmentation and increase the dataset by 3 folds.
4. Feature Extraction: The system should extract relevant features from preprocessed images, including
shape, texture, intensity, and spatial relationships. Feature extraction methods should capture key
characteristics indicative of brain tumors while minimizing information loss.
5. Model Selection and Development: Investigate different deep learning architectures and select an
appropriate architecture (e.g., U-Net, 3D Convolutional Neural Networks (CNNs), V-Net, DeepMedic,
Residual Networks (ResNet), DenseNet, Attention Mechanisms, YOLO ) for development.
6. Train & Test Data: Split dataset into 70% training and 30% testing dataset and train the model
accordingly.
7. Evaluation and Fine-tuning: Assess the model's performance using standard NER evaluation metrics
(e.g. Accuracy, F1-score, precision, recall, Receiver Operating Characteristic (ROC) Curve and Area
Under the Curve (AUC)) and fine-tune the model for improved accuracy.
8. Confusion Matrix: Create a confusion matrix table to describe the performance of the classification
model.
9. Real-time Detection: The system should provide real-time or near-real-time tumor detection
capabilities to support efficient clinical workflows. Detection speed should be optimized to minimize
processing time while maintaining high accuracy.
10. User Interface: NeuroScan must have an intuitive and user-friendly interface accessible to medical
imaging professionals. The interface should facilitate image input, visualization of detected tumors,
and access to diagnostic information. It should support interactive features for adjusting parameters,
reviewing results, and annotating findings.
Dataset:
https://drive.google.com/file/d/1CKi6MtrSwARHvBQgJGzutfkAhEb_kuQi/view?usp=sharing
*You must use your VU email id to access/download the dataset.
Tools:
 Python
 jupyter notebook
 Colab
 PyQt
 wxPython
 Tkinter
 Kivy
 PySimpleGUI
Prerequisite:
Artificial Intelligence, Machine Learning, and Image Processing, Computer Vision concepts,
“Admin (student) will cover a short course relevant to the mentioned concepts besides SRS and Design initial
documentation.”
Helping Material
Python
https://www.python.org/
https://www.w3schools.com/python/
https://www.tutorialspoint.com/python/index.htm
Deep Learning:
https://www.tutorialspoint.com/python_deep_learning/index.htm
https://www.tutorialspoint.com/deep-learning-tutorials/index.asp
Deep Learning Crash Course for Beginners (youtube.com)
Deep Learning | What is Deep Learning? | Deep Learning Tutorial For Beginners | 2023 | Simplilearn
(youtube.com)
Image Processing:
Python tutorials for image processing and machine learning - YouTube
Supervisor:
Name: Umair Ali
Email ID: umairali@vu.edu.pk
Skype ID: live:umairalihamid_1

Spring 2024 - CS619 - 9340

Uploaded by

Copyright:

Available Formats

Spring 2024 - CS619 - 9340

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Spring 2024 - CS619 - 9340

Uploaded by

Copyright:

Available Formats

NeuroScan: Advanced Brain Tumor Detection System

Project Domain / Category

Admin (Student) will perform all the following tasks.

You might also like