RSNA Intracranial Aneurysm Detection

Kaggle Competition URL

Explored the DICOM files using duckdb and polars
Extracted metadata from all the training DICOM series via multithreading
Appended the training metadata along with the 14 target variables using duckdb
Saved the final dataframe containing DICOM image metadata, coordinates of aneurysm location (for applicable images), training metadata in a single parquet file
Determined the bias in the given training data
Started to train a baseline model using transfer learning from Keras InceptionV3 model with inputs of size (128,128,3)

Name		Name	Last commit message	Last commit date
Latest commit History 57 Commits
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
Suman-Notebook-2.ipynb		Suman-Notebook-2.ipynb
baseline-and-feature-engineering.ipynb		baseline-and-feature-engineering.ipynb
rsna-submission.ipynb		rsna-submission.ipynb
suman-notebook-1.ipynb		suman-notebook-1.ipynb
using-tfrecord.ipynb		using-tfrecord.ipynb