Skip to content

ShaikhLuiza/PulmoNet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Lung Cancer Detection Using ResNet50 CNN

Overview

Lung cancer remains one of the leading causes of mortality worldwide, emphasizing the urgent need for early and accurate detection to improve patient outcomes. This project leverages Convolutional Neural Networks (CNNs), specifically the ResNet50 architecture, to detect lung cancer from Computed Tomography (CT) scans.

ResNet50, with its deep residual connections, excels at analyzing complex medical images and extracting critical features necessary for precise diagnosis. The project highlights meticulous data preparation and evaluates performance using metrics such as accuracy, sensitivity, and specificity.


Features

  • Deep Learning Model: Employs ResNet50 for robust image feature extraction.
  • Medical Image Analysis: Detects lung cancer from CT scan images.
  • High Accuracy: Optimized to achieve strong performance metrics.
  • Data Preprocessing: Handles normalization, resizing, and augmentation to improve model robustness.
  • Evaluation: Uses accuracy, sensitivity, specificity, and other key performance indicators.

Dataset

  • Source: Kaggle
  • Format: CT scan images in .png / .jpg.
  • Classes:
    • Benign
    • Malignant

Special attention is given to class imbalance to ensure accurate model predictions.


Technologies Used

  • Programming Language: Python
  • Deep Learning Framework: TensorFlow / Keras
  • Libraries: OpenCV, NumPy, Pandas, Matplotlib, Scikit-learn

Model Performance

  • Training Accuracy: ~90%
  • Testing Accuracy: ~70%

CNN-based approaches, particularly ResNet50, have shown strong performance in lung cancer detection, often achieving accuracies exceeding 90%.


Challenges & Future Work

  • Class Imbalance: Needs careful handling through augmentation and weighted loss.
  • Model Generalizability: Ensuring robustness across diverse patient populations and imaging conditions.
  • Data Limitations: Strategies to handle small or biased datasets to enhance reliability in real-world settings.

Keywords: Image Processing, Feature Extraction, Classification, Model Training, Cancer Detection, ResNet50

Releases

No releases published

Packages

 
 
 

Contributors

Languages