Problem Statement
Chest X-ray exam is one of the most frequent and cost-effective medical imaging examination. However clinical diagnosis of chest X-ray can be challenging, and sometimes believed to be harder than diagnosis via chest CT imaging. To achieve clinically relevant computer-aided detection and diagnosis (CAD) in real world medical sites on all data settings of chest X-rays is still very difficult, if not impossible when only several thousands of images are employed for study.
Started in 1953, National Institute of Health - Clinical Centre is one the leading hospitals in US. They are the active partners in medical discovery. Currently, there are around 1600 clinical research studies in progress at NIH centre, USA. With a support staff of around 620 nurses, in 2016, they handled more than 10,400 new patient
As a part of a research study to explore deep learning techniques, NIH has recently open-sourced their dataset of frontal chest X-ray images of patients.
Your task is to identify the class of thorax diseases from the given chest x-ray images.
Download Dataset (Images)
Download Dataset (csv files)
Torrent File
Join our slack channel to discuss ML and DL here.
Data Description
You are given two separate files to download (images and csv files). The train data has information for 18577 patients and test data has information for 12386 patients. The target variable has 14 types of thorax disease. Some part of data has been anonymised to restrict fraudulent submissions.
| Variable | Description |
|---|---|
| row_id | unique patient id |
| age | patient age |
| gender | patient gender |
| view position | position of image (binary) |
| image_name | x-ray image corresponding to patient |
| detected | target variable |
Submission Format
A participant has to submit a .csv file containing row_id and predictions as labels in detected. Check the sample submission file for correct format.
row_id, detected
id_100, class_1
id_10002, class_6
id_10005, class_8
id_10008, class_8
id_10009, class_7
Evaluation Metric
The submissions will be evaluated based on weighted F1 Score.
