DevMesh

Pneumothorax Detection using Computer Vision with CNN using Human Chest X-Ray Dataset

Mayank Tale

Mayank Tale

Delhi, Delhi

1 0
  • 0 Collaborators

We use the ChestXRay dataset, a set of X-ray images labeled with corresponding diagnoses to build a model that classifies patients with a Pneumothorax (ie. dropped or collapsed lung). ...learn more

Project status: Published/In Market

Artificial Intelligence

Intel Technologies
Other

Code Samples [1]Links [1]

Overview / Usage

The Chest X-Ray dataset consists of over 100,000 different X-rays from more than 30,000 patients labeled with different disease conditions. Our goal for this lab is to use a subset of the dataset to train a classifier that is able to accurately infer the presence of pneumothorax.But the Dataset is too Noisy!!

Pneumothorax is a condition that occurs when there are abnormal amounts of air in the space between the lung and chest wall. This reduces the capacity to which the lung is able to expand and fill with air, leading to oxygen shortage and low blood pressure. Lack of treatment may lead to worsening symptoms and even death.

Methodology / Approach

The idea would be to make a model sufficient to give us about 80% accuracy (as the dataset is too Noisy!!), Many Hidden Nets would be required. And Yes, GPU power would be required to train this model quickly.

Technologies Used

Keras, CNN(Computer Vision), Python, OpenCV

Repository

https://github.com/CodeLogist/Pneumothorax-Detection-CNN-Keras

Comments (0)